Hugging Face's logo
ziyan98 's Collections
E2K: experience to knowledge
E2K cd82 checkpoints: act_ctrl-cc (frozen-Adam attempt)
E2K cd82 checkpoints: sft_first_idd-cc (frozen-Adam attempt)
E2K cd82 checkpoints: sft_first-curatedL1
E2K cd82 checkpoints: sft_first
E2K cd82 checkpoints: baseline
E2K cd82 checkpoints: sft_joint-claimform
E2K cd82 checkpoints: sft_interleaved-claimform
E2K cd82 checkpoints: sft_first-claimform
E2K cd82 checkpoints: sft_joint-cc
E2K cd82 checkpoints: sft_first-cc
E2K cd82 checkpoints: sft_inter_decl-cc2
E2K cd82 checkpoints: sft_first_decl-cc2
E2K cd82 checkpoints: sft_first_decl-cc
E2K cd82 checkpoints: sft_first_armc-cc
E2K cd82 checkpoints: sft_first_decl-cc4
E2K cd82 checkpoints: act_ctrl-cc
E2K cd82 checkpoints: baseline-cc
E2K cd82 checkpoints: sft_first_idd-cc
E2K cd82 checkpoints: sft_first_armc-cc6

E2K: experience to knowledge

Entry point: 5 datasets (trajectories, verified knowledge, RL runs) + one checkpoint per RL run; each run has its own series collection.