├── .gitignore ├── .gitmodules ├── LICENSE ├── README.md ├── README_zh.md ├── easysteer ├── __init__.py ├── hidden_states │ ├── __init__.py │ └── capture.py ├── reft │ ├── basic_demo.py │ ├── pyreft │ │ ├── __init__.py │ │ ├── analysis │ │ │ ├── __init__.py │ │ │ └── visualization.py │ │ ├── core │ │ │ ├── __init__.py │ │ │ ├── base.py │ │ │ ├── interventions.py │ │ │ ├── modeling │ │ │ │ ├── __init__.py │ │ │ │ ├── backpack_gpt2 │ │ │ │ │ ├── __init__.py │ │ │ │ │ ├── modelings_backpack_gpt2.py │ │ │ │ │ └── modelings_intervenable_backpack_gpt2.py │ │ │ │ ├── basic_utils.py │ │ │ │ ├── blip │ │ │ │ │ ├── __init__.py │ │ │ │ │ ├── modelings_blip.py │ │ │ │ │ ├── modelings_blip_itm.py │ │ │ │ │ ├── modelings_intervenable_blip.py │ │ │ │ │ └── modelings_intervenable_blip_itm.py │ │ │ │ ├── common.py │ │ │ │ ├── configuration_intervenable_model.py │ │ │ │ ├── constants.py │ │ │ │ ├── esm │ │ │ │ │ ├── __init__.py │ │ │ │ │ └── modelings_intervenable_esm.py │ │ │ │ ├── gemma │ │ │ │ │ ├── __init__.py │ │ │ │ │ └── modelings_intervenable_gemma.py │ │ │ │ ├── gemma2 │ │ │ │ │ ├── __init__.py │ │ │ │ │ └── modelings_intervenable_gemma2.py │ │ │ │ ├── gpt2 │ │ │ │ │ ├── __init__.py │ │ │ │ │ └── modelings_intervenable_gpt2.py │ │ │ │ ├── gpt_neo │ │ │ │ │ ├── __init__.py │ │ │ │ │ └── modelings_intervenable_gpt_neo.py │ │ │ │ ├── gpt_neox │ │ │ │ │ ├── __init__.py │ │ │ │ │ └── modelings_intervenable_gpt_neox.py │ │ │ │ ├── gru │ │ │ │ │ ├── __init__.py │ │ │ │ │ ├── modelings_gru.py │ │ │ │ │ └── modelings_intervenable_gru.py │ │ │ │ ├── intervenable_base.py │ │ │ │ ├── intervenable_modelcard.py │ │ │ │ ├── intervention_utils.py │ │ │ │ ├── interventions.py │ │ │ │ ├── layers.py │ │ │ │ ├── llama │ │ │ │ │ ├── __init__.py │ │ │ │ │ └── modelings_intervenable_llama.py │ │ │ │ ├── llava │ │ │ │ │ ├── __init__.py │ │ │ │ │ └── modelings_intervenable_llava.py │ │ │ │ ├── mistral │ │ │ │ │ ├── __init__.py │ │ │ │ │ └── modellings_intervenable_mistral.py │ │ │ │ ├── mlp │ │ │ │ │ ├── __init__.py │ │ │ │ │ ├── modelings_intervenable_mlp.py │ │ │ │ │ └── modelings_mlp.py │ │ │ │ ├── modeling_utils.py │ │ │ │ ├── olmo │ │ │ │ │ ├── __init__.py │ │ │ │ │ └── modelings_intervenable_olmo.py │ │ │ │ └── qwen2 │ │ │ │ │ ├── __init__.py │ │ │ │ │ └── modelings_intervenable_qwen2.py │ │ │ └── utils.py │ │ ├── data │ │ │ ├── __init__.py │ │ │ ├── causal_model.py │ │ │ └── dataset.py │ │ ├── examples │ │ │ ├── __init__.py │ │ │ ├── basic_demo.py │ │ │ └── notebooks │ │ │ │ └── tutorial.ipynb │ │ ├── reft │ │ │ ├── __init__.py │ │ │ ├── algorithms │ │ │ │ ├── __init__.py │ │ │ │ ├── bias.py │ │ │ │ ├── consreft.py │ │ │ │ ├── direft.py │ │ │ │ ├── lobireft.py │ │ │ │ ├── loreft.py │ │ │ │ ├── nodireft.py │ │ │ │ ├── noreft.py │ │ │ │ └── utils.py │ │ │ ├── config.py │ │ │ ├── model.py │ │ │ ├── trainer.py │ │ │ └── utils.py │ │ ├── requirements.txt │ │ ├── setup.py │ │ ├── test_new_architecture.py │ │ └── tests │ │ │ └── __init__.py │ ├── results │ │ ├── loreft │ │ │ ├── config.json │ │ │ └── intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin │ │ └── ssv │ │ │ ├── config.json │ │ │ └── intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin │ ├── ssv.py │ └── test_demo.py └── steer │ ├── __init__.py │ ├── diffmean.py │ ├── lat.py │ ├── linear_probe.py │ ├── pca.py │ ├── sae.py │ ├── unified_interface.py │ └── utils.py ├── experiment ├── efficiency │ ├── easysteer_multi.ipynb │ ├── easysteer_single.ipynb │ ├── pyreft.ipynb │ ├── pyreft_batch.ipynb │ ├── repeng.ipynb │ ├── repeng_batch.ipynb │ └── vllm_baseline.ipynb ├── hallucination │ ├── TruthfulQA.csv │ ├── baseline.ipynb │ ├── data.ipynb │ ├── eval.ipynb │ ├── mc_task.json │ └── steer.ipynb └── math │ ├── baseline.ipynb │ ├── data_construction.ipynb │ ├── execution_avg_vector.gguf │ ├── reflection_avg_vector.gguf │ ├── steer.ipynb │ └── transition_avg_vector.gguf ├── figures ├── app.png ├── arch.png ├── logo.png ├── speed.png └── wechat.png ├── frontend ├── README.md ├── app.py ├── chat_api.py ├── configs │ ├── chat │ │ ├── cat_mode.json │ │ ├── chinese_mode.json │ │ ├── happy_mode.json │ │ └── reject_mode.json │ ├── extraction │ │ ├── emotion_diffmean.json │ │ ├── emotion_lat.json │ │ ├── emotion_pca.json │ │ └── personality_diffmean.json │ ├── inference │ │ ├── emoji_loreft.json │ │ └── emotion_direct.json │ ├── multi_vector │ │ └── refusal_direction.json │ ├── training │ │ ├── emoji_bias.json │ │ ├── emoji_loreft.json │ │ └── emotion_loreft.json │ ├── training_emoji_bias.json │ ├── training_emoji_loreft.json │ └── training_emotion_loreft.json ├── demo_training.py ├── extraction_api.py ├── i18n.js ├── index.html ├── inference_api.py ├── requirements.txt ├── results │ ├── emoji_bias │ │ ├── config.json │ │ └── intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin │ ├── emoji_loreft │ │ ├── config.json │ │ └── intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin │ └── emotion_loreft │ │ ├── config.json │ │ └── intkey_layer_12_comp_attention_output_unit_pos_nunit_1#0.bin ├── sae_api.py ├── start.bat ├── start.sh ├── start_server.py ├── static │ ├── css │ │ ├── base.css │ │ ├── chat.css │ │ ├── extraction.css │ │ ├── inference.css │ │ ├── main.css │ │ ├── responsive.css │ │ ├── sae-explore.css │ │ └── training.css │ ├── js │ │ ├── chat.js │ │ ├── extraction.js │ │ ├── inference.js │ │ ├── main.js │ │ ├── multi-vector.js │ │ ├── sae-explore.js │ │ ├── training.js │ │ ├── ui.js │ │ └── utils.js │ └── templates │ │ ├── chat.html │ │ ├── extraction.html │ │ ├── inference.html │ │ └── training.html └── training_api.py ├── pyproject.toml ├── replications ├── bipo │ ├── README.md │ ├── power_seeking_steer.ipynb │ └── vec_ep20_layer15.pt ├── cast │ ├── README.md │ ├── alpaca.json │ ├── data_construction.ipynb │ ├── refusal_direction_steer.ipynb │ └── refuse-pca.gguf ├── controlingthinkingspeed │ ├── MATH500.gguf │ ├── README.md │ ├── data_construction.ipynb │ └── steer.ipynb ├── creative_writing │ ├── README.md │ ├── create.gguf │ ├── creative_steer.ipynb │ └── data_construction.ipynb ├── fractreason │ ├── MATH500.gguf │ ├── README.md │ ├── data_construction.ipynb │ └── fractreason_steer.ipynb ├── improve_reasoning │ ├── README.md │ ├── data.ipynb │ ├── reason.gguf │ └── steer.ipynb ├── lm_steer │ ├── README.md │ └── lm_steer.ipynb ├── loreft │ ├── README.md │ ├── loreft_steer.ipynb │ ├── loreft_train.ipynb │ └── weight │ │ ├── config.json │ │ └── intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin ├── refusal_direction │ ├── README.md │ ├── data_construction.ipynb │ ├── diffmean-1.gguf │ ├── diffmean-2.gguf │ ├── diffmean-3.gguf │ ├── diffmean-4.gguf │ └── refusal_direction_steer.ipynb ├── sae_entities │ ├── README.md │ ├── james.pt │ └── steer.ipynb ├── sake │ ├── README.md │ ├── data_construction.ipynb │ └── sake_steer.ipynb ├── seal │ ├── README.md │ ├── data_construction.ipynb │ ├── execution_avg_vector.gguf │ ├── reflection_avg_vector.gguf │ ├── seal_steer.ipynb │ └── transition_avg_vector.gguf ├── sharp │ ├── README.md │ ├── sharp.ipynb │ ├── task_vector_layer-10.npy │ ├── task_vector_layer-10.pt │ └── test.jpg └── steerable_chatbot │ ├── README.md │ ├── data.ipynb │ ├── steer.ipynb │ └── style-probe.gguf ├── setup.py └── vectors └── happy_diffmean.gguf /.gitignore: -------------------------------------------------------------------------------- 1 | __pycache__/ 2 | .idea/ 3 | /temp/ 4 | easysteer.egg-info/ -------------------------------------------------------------------------------- /.gitmodules: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/.gitmodules -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/LICENSE -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/README.md -------------------------------------------------------------------------------- /README_zh.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/README_zh.md -------------------------------------------------------------------------------- /easysteer/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/__init__.py -------------------------------------------------------------------------------- /easysteer/hidden_states/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/hidden_states/__init__.py -------------------------------------------------------------------------------- /easysteer/hidden_states/capture.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/hidden_states/capture.py -------------------------------------------------------------------------------- /easysteer/reft/basic_demo.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/basic_demo.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/analysis/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/analysis/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/analysis/visualization.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/analysis/visualization.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/base.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/base.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/interventions.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/interventions.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/backpack_gpt2/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/backpack_gpt2/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/backpack_gpt2/modelings_backpack_gpt2.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/backpack_gpt2/modelings_backpack_gpt2.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/backpack_gpt2/modelings_intervenable_backpack_gpt2.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/backpack_gpt2/modelings_intervenable_backpack_gpt2.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/basic_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/basic_utils.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/blip/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/blip/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/blip/modelings_blip.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/blip/modelings_blip.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/blip/modelings_blip_itm.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/blip/modelings_blip_itm.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/blip/modelings_intervenable_blip.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/blip/modelings_intervenable_blip.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/blip/modelings_intervenable_blip_itm.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/blip/modelings_intervenable_blip_itm.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/common.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/common.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/configuration_intervenable_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/configuration_intervenable_model.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/constants.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/constants.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/esm/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/esm/modelings_intervenable_esm.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/esm/modelings_intervenable_esm.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/gemma/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/gemma/modelings_intervenable_gemma.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/gemma/modelings_intervenable_gemma.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/gemma2/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/gemma2/modelings_intervenable_gemma2.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/gemma2/modelings_intervenable_gemma2.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/gpt2/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/gpt2/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/gpt2/modelings_intervenable_gpt2.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/gpt2/modelings_intervenable_gpt2.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/gpt_neo/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/gpt_neo/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/gpt_neo/modelings_intervenable_gpt_neo.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/gpt_neo/modelings_intervenable_gpt_neo.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/gpt_neox/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/gpt_neox/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/gpt_neox/modelings_intervenable_gpt_neox.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/gpt_neox/modelings_intervenable_gpt_neox.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/gru/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/gru/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/gru/modelings_gru.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/gru/modelings_gru.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/gru/modelings_intervenable_gru.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/gru/modelings_intervenable_gru.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/intervenable_base.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/intervenable_base.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/intervenable_modelcard.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/intervenable_modelcard.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/intervention_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/intervention_utils.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/interventions.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/interventions.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/layers.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/layers.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/llama/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/llama/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/llama/modelings_intervenable_llama.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/llama/modelings_intervenable_llama.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/llava/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/llava/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/llava/modelings_intervenable_llava.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/llava/modelings_intervenable_llava.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/mistral/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/mistral/modellings_intervenable_mistral.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/mistral/modellings_intervenable_mistral.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/mlp/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/mlp/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/mlp/modelings_intervenable_mlp.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/mlp/modelings_intervenable_mlp.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/mlp/modelings_mlp.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/mlp/modelings_mlp.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/modeling_utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/modeling_utils.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/olmo/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/olmo/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/olmo/modelings_intervenable_olmo.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/olmo/modelings_intervenable_olmo.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/qwen2/__init__.py: -------------------------------------------------------------------------------- 1 | -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/modeling/qwen2/modelings_intervenable_qwen2.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/modeling/qwen2/modelings_intervenable_qwen2.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/core/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/core/utils.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/data/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/data/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/data/causal_model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/data/causal_model.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/data/dataset.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/data/dataset.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/examples/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/examples/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/examples/basic_demo.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/examples/basic_demo.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/examples/notebooks/tutorial.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/examples/notebooks/tutorial.ipynb -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/algorithms/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/algorithms/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/algorithms/bias.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/algorithms/bias.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/algorithms/consreft.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/algorithms/consreft.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/algorithms/direft.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/algorithms/direft.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/algorithms/lobireft.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/algorithms/lobireft.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/algorithms/loreft.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/algorithms/loreft.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/algorithms/nodireft.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/algorithms/nodireft.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/algorithms/noreft.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/algorithms/noreft.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/algorithms/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/algorithms/utils.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/config.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/config.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/model.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/model.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/trainer.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/trainer.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/reft/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/reft/utils.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/requirements.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/requirements.txt -------------------------------------------------------------------------------- /easysteer/reft/pyreft/setup.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/setup.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/test_new_architecture.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/test_new_architecture.py -------------------------------------------------------------------------------- /easysteer/reft/pyreft/tests/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/pyreft/tests/__init__.py -------------------------------------------------------------------------------- /easysteer/reft/results/loreft/config.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/results/loreft/config.json -------------------------------------------------------------------------------- /easysteer/reft/results/loreft/intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/results/loreft/intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin -------------------------------------------------------------------------------- /easysteer/reft/results/ssv/config.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/results/ssv/config.json -------------------------------------------------------------------------------- /easysteer/reft/results/ssv/intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/results/ssv/intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin -------------------------------------------------------------------------------- /easysteer/reft/ssv.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/ssv.py -------------------------------------------------------------------------------- /easysteer/reft/test_demo.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/reft/test_demo.py -------------------------------------------------------------------------------- /easysteer/steer/__init__.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/steer/__init__.py -------------------------------------------------------------------------------- /easysteer/steer/diffmean.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/steer/diffmean.py -------------------------------------------------------------------------------- /easysteer/steer/lat.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/steer/lat.py -------------------------------------------------------------------------------- /easysteer/steer/linear_probe.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/steer/linear_probe.py -------------------------------------------------------------------------------- /easysteer/steer/pca.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/steer/pca.py -------------------------------------------------------------------------------- /easysteer/steer/sae.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/steer/sae.py -------------------------------------------------------------------------------- /easysteer/steer/unified_interface.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/steer/unified_interface.py -------------------------------------------------------------------------------- /easysteer/steer/utils.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/easysteer/steer/utils.py -------------------------------------------------------------------------------- /experiment/efficiency/easysteer_multi.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/efficiency/easysteer_multi.ipynb -------------------------------------------------------------------------------- /experiment/efficiency/easysteer_single.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/efficiency/easysteer_single.ipynb -------------------------------------------------------------------------------- /experiment/efficiency/pyreft.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/efficiency/pyreft.ipynb -------------------------------------------------------------------------------- /experiment/efficiency/pyreft_batch.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/efficiency/pyreft_batch.ipynb -------------------------------------------------------------------------------- /experiment/efficiency/repeng.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/efficiency/repeng.ipynb -------------------------------------------------------------------------------- /experiment/efficiency/repeng_batch.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/efficiency/repeng_batch.ipynb -------------------------------------------------------------------------------- /experiment/efficiency/vllm_baseline.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/efficiency/vllm_baseline.ipynb -------------------------------------------------------------------------------- /experiment/hallucination/TruthfulQA.csv: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/hallucination/TruthfulQA.csv -------------------------------------------------------------------------------- /experiment/hallucination/baseline.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/hallucination/baseline.ipynb -------------------------------------------------------------------------------- /experiment/hallucination/data.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/hallucination/data.ipynb -------------------------------------------------------------------------------- /experiment/hallucination/eval.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/hallucination/eval.ipynb -------------------------------------------------------------------------------- /experiment/hallucination/mc_task.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/hallucination/mc_task.json -------------------------------------------------------------------------------- /experiment/hallucination/steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/hallucination/steer.ipynb -------------------------------------------------------------------------------- /experiment/math/baseline.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/math/baseline.ipynb -------------------------------------------------------------------------------- /experiment/math/data_construction.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/math/data_construction.ipynb -------------------------------------------------------------------------------- /experiment/math/execution_avg_vector.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/math/execution_avg_vector.gguf -------------------------------------------------------------------------------- /experiment/math/reflection_avg_vector.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/math/reflection_avg_vector.gguf -------------------------------------------------------------------------------- /experiment/math/steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/math/steer.ipynb -------------------------------------------------------------------------------- /experiment/math/transition_avg_vector.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/experiment/math/transition_avg_vector.gguf -------------------------------------------------------------------------------- /figures/app.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/figures/app.png -------------------------------------------------------------------------------- /figures/arch.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/figures/arch.png -------------------------------------------------------------------------------- /figures/logo.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/figures/logo.png -------------------------------------------------------------------------------- /figures/speed.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/figures/speed.png -------------------------------------------------------------------------------- /figures/wechat.png: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/figures/wechat.png -------------------------------------------------------------------------------- /frontend/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/README.md -------------------------------------------------------------------------------- /frontend/app.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/app.py -------------------------------------------------------------------------------- /frontend/chat_api.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/chat_api.py -------------------------------------------------------------------------------- /frontend/configs/chat/cat_mode.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/chat/cat_mode.json -------------------------------------------------------------------------------- /frontend/configs/chat/chinese_mode.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/chat/chinese_mode.json -------------------------------------------------------------------------------- /frontend/configs/chat/happy_mode.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/chat/happy_mode.json -------------------------------------------------------------------------------- /frontend/configs/chat/reject_mode.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/chat/reject_mode.json -------------------------------------------------------------------------------- /frontend/configs/extraction/emotion_diffmean.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/extraction/emotion_diffmean.json -------------------------------------------------------------------------------- /frontend/configs/extraction/emotion_lat.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/extraction/emotion_lat.json -------------------------------------------------------------------------------- /frontend/configs/extraction/emotion_pca.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/extraction/emotion_pca.json -------------------------------------------------------------------------------- /frontend/configs/extraction/personality_diffmean.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/extraction/personality_diffmean.json -------------------------------------------------------------------------------- /frontend/configs/inference/emoji_loreft.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/inference/emoji_loreft.json -------------------------------------------------------------------------------- /frontend/configs/inference/emotion_direct.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/inference/emotion_direct.json -------------------------------------------------------------------------------- /frontend/configs/multi_vector/refusal_direction.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/multi_vector/refusal_direction.json -------------------------------------------------------------------------------- /frontend/configs/training/emoji_bias.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/training/emoji_bias.json -------------------------------------------------------------------------------- /frontend/configs/training/emoji_loreft.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/training/emoji_loreft.json -------------------------------------------------------------------------------- /frontend/configs/training/emotion_loreft.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/training/emotion_loreft.json -------------------------------------------------------------------------------- /frontend/configs/training_emoji_bias.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/training_emoji_bias.json -------------------------------------------------------------------------------- /frontend/configs/training_emoji_loreft.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/training_emoji_loreft.json -------------------------------------------------------------------------------- /frontend/configs/training_emotion_loreft.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/configs/training_emotion_loreft.json -------------------------------------------------------------------------------- /frontend/demo_training.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/demo_training.py -------------------------------------------------------------------------------- /frontend/extraction_api.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/extraction_api.py -------------------------------------------------------------------------------- /frontend/i18n.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/i18n.js -------------------------------------------------------------------------------- /frontend/index.html: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/index.html -------------------------------------------------------------------------------- /frontend/inference_api.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/inference_api.py -------------------------------------------------------------------------------- /frontend/requirements.txt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/requirements.txt -------------------------------------------------------------------------------- /frontend/results/emoji_bias/config.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/results/emoji_bias/config.json -------------------------------------------------------------------------------- /frontend/results/emoji_bias/intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/results/emoji_bias/intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin -------------------------------------------------------------------------------- /frontend/results/emoji_loreft/config.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/results/emoji_loreft/config.json -------------------------------------------------------------------------------- /frontend/results/emoji_loreft/intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/results/emoji_loreft/intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin -------------------------------------------------------------------------------- /frontend/results/emotion_loreft/config.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/results/emotion_loreft/config.json -------------------------------------------------------------------------------- /frontend/results/emotion_loreft/intkey_layer_12_comp_attention_output_unit_pos_nunit_1#0.bin: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/results/emotion_loreft/intkey_layer_12_comp_attention_output_unit_pos_nunit_1#0.bin -------------------------------------------------------------------------------- /frontend/sae_api.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/sae_api.py -------------------------------------------------------------------------------- /frontend/start.bat: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/start.bat -------------------------------------------------------------------------------- /frontend/start.sh: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/start.sh -------------------------------------------------------------------------------- /frontend/start_server.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/start_server.py -------------------------------------------------------------------------------- /frontend/static/css/base.css: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/css/base.css -------------------------------------------------------------------------------- /frontend/static/css/chat.css: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/css/chat.css -------------------------------------------------------------------------------- /frontend/static/css/extraction.css: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/css/extraction.css -------------------------------------------------------------------------------- /frontend/static/css/inference.css: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/css/inference.css -------------------------------------------------------------------------------- /frontend/static/css/main.css: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/css/main.css -------------------------------------------------------------------------------- /frontend/static/css/responsive.css: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/css/responsive.css -------------------------------------------------------------------------------- /frontend/static/css/sae-explore.css: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/css/sae-explore.css -------------------------------------------------------------------------------- /frontend/static/css/training.css: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/css/training.css -------------------------------------------------------------------------------- /frontend/static/js/chat.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/js/chat.js -------------------------------------------------------------------------------- /frontend/static/js/extraction.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/js/extraction.js -------------------------------------------------------------------------------- /frontend/static/js/inference.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/js/inference.js -------------------------------------------------------------------------------- /frontend/static/js/main.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/js/main.js -------------------------------------------------------------------------------- /frontend/static/js/multi-vector.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/js/multi-vector.js -------------------------------------------------------------------------------- /frontend/static/js/sae-explore.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/js/sae-explore.js -------------------------------------------------------------------------------- /frontend/static/js/training.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/js/training.js -------------------------------------------------------------------------------- /frontend/static/js/ui.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/js/ui.js -------------------------------------------------------------------------------- /frontend/static/js/utils.js: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/js/utils.js -------------------------------------------------------------------------------- /frontend/static/templates/chat.html: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/templates/chat.html -------------------------------------------------------------------------------- /frontend/static/templates/extraction.html: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/templates/extraction.html -------------------------------------------------------------------------------- /frontend/static/templates/inference.html: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/templates/inference.html -------------------------------------------------------------------------------- /frontend/static/templates/training.html: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/static/templates/training.html -------------------------------------------------------------------------------- /frontend/training_api.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/frontend/training_api.py -------------------------------------------------------------------------------- /pyproject.toml: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/pyproject.toml -------------------------------------------------------------------------------- /replications/bipo/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/bipo/README.md -------------------------------------------------------------------------------- /replications/bipo/power_seeking_steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/bipo/power_seeking_steer.ipynb -------------------------------------------------------------------------------- /replications/bipo/vec_ep20_layer15.pt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/bipo/vec_ep20_layer15.pt -------------------------------------------------------------------------------- /replications/cast/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/cast/README.md -------------------------------------------------------------------------------- /replications/cast/alpaca.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/cast/alpaca.json -------------------------------------------------------------------------------- /replications/cast/data_construction.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/cast/data_construction.ipynb -------------------------------------------------------------------------------- /replications/cast/refusal_direction_steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/cast/refusal_direction_steer.ipynb -------------------------------------------------------------------------------- /replications/cast/refuse-pca.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/cast/refuse-pca.gguf -------------------------------------------------------------------------------- /replications/controlingthinkingspeed/MATH500.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/controlingthinkingspeed/MATH500.gguf -------------------------------------------------------------------------------- /replications/controlingthinkingspeed/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/controlingthinkingspeed/README.md -------------------------------------------------------------------------------- /replications/controlingthinkingspeed/data_construction.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/controlingthinkingspeed/data_construction.ipynb -------------------------------------------------------------------------------- /replications/controlingthinkingspeed/steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/controlingthinkingspeed/steer.ipynb -------------------------------------------------------------------------------- /replications/creative_writing/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/creative_writing/README.md -------------------------------------------------------------------------------- /replications/creative_writing/create.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/creative_writing/create.gguf -------------------------------------------------------------------------------- /replications/creative_writing/creative_steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/creative_writing/creative_steer.ipynb -------------------------------------------------------------------------------- /replications/creative_writing/data_construction.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/creative_writing/data_construction.ipynb -------------------------------------------------------------------------------- /replications/fractreason/MATH500.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/fractreason/MATH500.gguf -------------------------------------------------------------------------------- /replications/fractreason/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/fractreason/README.md -------------------------------------------------------------------------------- /replications/fractreason/data_construction.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/fractreason/data_construction.ipynb -------------------------------------------------------------------------------- /replications/fractreason/fractreason_steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/fractreason/fractreason_steer.ipynb -------------------------------------------------------------------------------- /replications/improve_reasoning/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/improve_reasoning/README.md -------------------------------------------------------------------------------- /replications/improve_reasoning/data.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/improve_reasoning/data.ipynb -------------------------------------------------------------------------------- /replications/improve_reasoning/reason.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/improve_reasoning/reason.gguf -------------------------------------------------------------------------------- /replications/improve_reasoning/steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/improve_reasoning/steer.ipynb -------------------------------------------------------------------------------- /replications/lm_steer/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/lm_steer/README.md -------------------------------------------------------------------------------- /replications/lm_steer/lm_steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/lm_steer/lm_steer.ipynb -------------------------------------------------------------------------------- /replications/loreft/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/loreft/README.md -------------------------------------------------------------------------------- /replications/loreft/loreft_steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/loreft/loreft_steer.ipynb -------------------------------------------------------------------------------- /replications/loreft/loreft_train.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/loreft/loreft_train.ipynb -------------------------------------------------------------------------------- /replications/loreft/weight/config.json: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/loreft/weight/config.json -------------------------------------------------------------------------------- /replications/loreft/weight/intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/loreft/weight/intkey_layer_8_comp_block_output_unit_pos_nunit_1#0.bin -------------------------------------------------------------------------------- /replications/refusal_direction/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/refusal_direction/README.md -------------------------------------------------------------------------------- /replications/refusal_direction/data_construction.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/refusal_direction/data_construction.ipynb -------------------------------------------------------------------------------- /replications/refusal_direction/diffmean-1.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/refusal_direction/diffmean-1.gguf -------------------------------------------------------------------------------- /replications/refusal_direction/diffmean-2.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/refusal_direction/diffmean-2.gguf -------------------------------------------------------------------------------- /replications/refusal_direction/diffmean-3.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/refusal_direction/diffmean-3.gguf -------------------------------------------------------------------------------- /replications/refusal_direction/diffmean-4.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/refusal_direction/diffmean-4.gguf -------------------------------------------------------------------------------- /replications/refusal_direction/refusal_direction_steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/refusal_direction/refusal_direction_steer.ipynb -------------------------------------------------------------------------------- /replications/sae_entities/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/sae_entities/README.md -------------------------------------------------------------------------------- /replications/sae_entities/james.pt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/sae_entities/james.pt -------------------------------------------------------------------------------- /replications/sae_entities/steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/sae_entities/steer.ipynb -------------------------------------------------------------------------------- /replications/sake/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/sake/README.md -------------------------------------------------------------------------------- /replications/sake/data_construction.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/sake/data_construction.ipynb -------------------------------------------------------------------------------- /replications/sake/sake_steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/sake/sake_steer.ipynb -------------------------------------------------------------------------------- /replications/seal/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/seal/README.md -------------------------------------------------------------------------------- /replications/seal/data_construction.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/seal/data_construction.ipynb -------------------------------------------------------------------------------- /replications/seal/execution_avg_vector.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/seal/execution_avg_vector.gguf -------------------------------------------------------------------------------- /replications/seal/reflection_avg_vector.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/seal/reflection_avg_vector.gguf -------------------------------------------------------------------------------- /replications/seal/seal_steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/seal/seal_steer.ipynb -------------------------------------------------------------------------------- /replications/seal/transition_avg_vector.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/seal/transition_avg_vector.gguf -------------------------------------------------------------------------------- /replications/sharp/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/sharp/README.md -------------------------------------------------------------------------------- /replications/sharp/sharp.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/sharp/sharp.ipynb -------------------------------------------------------------------------------- /replications/sharp/task_vector_layer-10.npy: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/sharp/task_vector_layer-10.npy -------------------------------------------------------------------------------- /replications/sharp/task_vector_layer-10.pt: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/sharp/task_vector_layer-10.pt -------------------------------------------------------------------------------- /replications/sharp/test.jpg: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/sharp/test.jpg -------------------------------------------------------------------------------- /replications/steerable_chatbot/README.md: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/steerable_chatbot/README.md -------------------------------------------------------------------------------- /replications/steerable_chatbot/data.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/steerable_chatbot/data.ipynb -------------------------------------------------------------------------------- /replications/steerable_chatbot/steer.ipynb: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/steerable_chatbot/steer.ipynb -------------------------------------------------------------------------------- /replications/steerable_chatbot/style-probe.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/replications/steerable_chatbot/style-probe.gguf -------------------------------------------------------------------------------- /setup.py: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/setup.py -------------------------------------------------------------------------------- /vectors/happy_diffmean.gguf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/ZJU-REAL/EasySteer/HEAD/vectors/happy_diffmean.gguf --------------------------------------------------------------------------------