models
Open weights, fine-tunes and adapters. Every listing here comes live from the Hugging Face Hub, attributed to it, and links back to the source.
dpo-harmless_saferlhfhelpful_harmless_honestRM-HH-Mix_harmless_gpt3_20000_gemma2b_shuffleTrue_extractchosenTrueREPROD_dpo_harmlessharmless_gpt4_subset-1_modelgemma7b_maxsteps10000_bz8_lr5e-06RM-HH-Gemma_harmless_gpt3_20000_gemma2b_shuffleTrue_extractchosenTrueRM-HH-Mix_harmless_gpt3_20000_gemma2b_shuffleFalse_extractchosenFalsedpo_harmlessharmless_gamma0.0_beta0.1_subset20000_modelmistral7b_maxsteps5000_bz8_lr1e-05dpo_harmlessharmless_gpt3_gamma0.0_beta0.1_subset20000_modelmistral7b_maxsteps5000_bz8_lr5e-06dpo_harmlessharmless_gpt3_gamma0.0_beta0.1_subset20000_modelmistral7b_maxsteps5000_bz8_lr1e-05RM-HH-Gemma_harmless_gpt3_20000_gemma2b_shuffleFalse_extractchosenTrueRM-HH-AllMix_harmless_gpt3_20000_gemma2b_shuffleFalse_extractchosenTrueRM-harmless_harmless_contrast_loraR64_20000_gemma2b_lr5e-05_bs2_g4RM-HH-Gemma_harmless_gpt3_20000_gemma2b_shuffleFalse_extractchosenFalseRM-HH-AllMix_harmless_gpt3_20000_gemma2b_shuffleTrue_extractchosenFalsedpo_harmlessharmless_human_gamma0.0_beta0.1_subset-1_modelmistral7b_maxsteps5000_bz8_lr1e-06dpo_harmlessharmless_human_gamma100.0_beta0.1_subset-1_modelmistral7b_maxsteps5000_bz8_lr1e-06dpo_harmlessharmless_human_gamma5.0_beta0.1_subset-1_modelmistral7b_maxsteps5000_bz8_lr1e-06dpo_harmlessharmless_human_gamma1.0_beta0.1_subset-1_modelmistral7b_maxsteps5000_bz8_lr1e-06dpo_harmlessharmless_human_gamma30.0_beta0.1_subset-1_modelmistral7b_maxsteps5000_bz8_lr1e-06dpo_harmlessharmless_gpt4_subset20000_modelgpt2_maxsteps5000_bz8_lr5e-06dpo-mistral-7b-harmless-v2dpo_harmlessharmless_gpt4_subset20000_modelgpt2_maxsteps5000_bz8_lr1e-05dpo_harmlessharmless_human_subset20000_modelgpt2_maxsteps5000_bz8_lr5e-06RM-harmless_harmless_contrast_loraR64_20000_gemma2b_lr1e-06_bs2_g4RM-harmless_harmless_human_loraR64_20000_gemma2b_lr5e-05_bs2_g4RM-harmless_harmless_gpt4_loraR64_20000_gemma2b_lr1e-05_bs2_g4REPROD_dpo_harmlessharmless_human_subset-1_modelgemma7b_maxsteps10000_bz8_lr5e-06RM-HH-AllMix_harmless_gpt3_20000_gemma2b_shuffleTrue_extractchosenTruedpo_harmlessharmless_human_subset20000_modelgpt2_maxsteps5000_bz8_lr1e-05dpo_harmlessharmless_gpt3_subset20000_modelgpt2_maxsteps5000_bz8_lr1e-05
