359 GB

Ctrl+K

1 contributor

History: 645 commits

karantonis

Back up retrained Gemma-2-2B sample_rand_C1 (no final soft-cap, phase2 step 400)

1872095 verified 1 day ago

analysis
MATH-train base labels (lane PL) 17 days ago
epd-lambda0
epd-lambda0 sample_rand_C2 phase2 3 months ago
il-judge-phase4-haiku
Phase 4 judge imitation: final policy and BC ablation 21 days ago
judge-imitation
judge-imitation: artem9/rl-uhead-gen-workdir 11 days ago
qwen3_5_4b_base
Upload folder using huggingface_hub 5 months ago
repl-gemma2-2b-nocap
Back up retrained Gemma-2-2B sample_rand_C1 (no final soft-cap, phase2 step 400) 1 day ago
repl-gemma2-2b
repl-gemma2-2b: epd_answeronly_matched C2 phase1_step500 (budget-matched world-only, rebuttal) 3 days ago
repl-gemma3-1b
repl-gemma3-1b: epd_answeronly_matched C2 phase1_step500 (budget-matched world-only, rebuttal) 3 days ago
repl-gemma3-4b
repl-gemma3-4b: epd_answeronly_matched C2 phase1_step500 (budget-matched world-only, rebuttal) 3 days ago
repl-granite-2b
repl-granite-2b: epd_answeronly_matched C2 phase1_step500 (budget-matched world-only, rebuttal) 3 days ago
repl-olmo2-1b
repl-olmo2-1b: epd_answeronly_matched C2 phase1_step500 (budget-matched world-only, rebuttal) 3 days ago
repl-olmo2-7b
repl-olmo2-7b: epd_answeronly_matched C2 phase1_step500 (budget-matched world-only, rebuttal) 3 days ago
repl-qwen3-4b
repl-qwen3-4b: epd_answeronly_matched C2 phase1_step500 (budget-matched world-only, rebuttal) 3 days ago
repl-qwen35-2b
repl-qwen35-2b: epd_answeronly_matched C2 phase1_step500 (budget-matched world-only, rebuttal) 3 days ago
repl-qwen35-4b
repl-qwen35-4b: epd_answeronly_matched C1 phase1_step500 (budget-matched world-only, rebuttal) 3 days ago
repl-qwen35-9b
repl-qwen35-9b: epd_answeronly_matched C2 phase1_step500 (budget-matched world-only, rebuttal) 3 days ago
rl-matched-normal-step400
4B matched control (fixed-normal prior, identical SFT+RL recipe to hidden); folded LoRA, step 400 3 months ago
rl-randprior-hidden-step400
Upload folder using huggingface_hub 5 months ago
rl-randprior-hidden
ckpt rl-randprior-hidden/global_step_25 5 months ago
rl-randprior-told-step400
randprior told RL final (step400) merged HF model 5 months ago
rl-randprior-told
ckpt rl-randprior-told/global_step_325 5 months ago
rl-scalar-hidden-step400
scalar_map RL arm (RLCR-style point-target reward, matched to hidden, step400) 4 months ago
rl_9b_hidden_lr3e5_step90
Upload folder using huggingface_hub 4 months ago
rl_lr1e5_step400
Upload folder using huggingface_hub 5 months ago
sft-randprior-hidden
prior-fix SFT (randprior-hidden) step_3000 adapter 5 months ago
sft-randprior-told
prior-fix SFT (randprior-told) step_3000 adapter 5 months ago
sft_9b_matched_normal
ckpt sft_9b_matched_normal/final 4 months ago
sft_9b_randprior_hidden
ckpt sft_9b_randprior_hidden/final 4 months ago
sft_final
Upload folder using huggingface_hub 5 months ago
sft_randprior_hidden_L2_step1750
Upload folder using huggingface_hub 3 months ago
truthrl-inenv-olmo2-1b
in-env TruthRL ablation: README 22 days ago
verbalize-randprior-hidden-v1
Upload folder using huggingface_hub 5 months ago
verbalize-randprior-told-v1
Upload folder using huggingface_hub 5 months ago
verbalize_i4_v1
Stage-3 verbalization adapter v1 (LoRA r64, self-distil to frozen-I4 belief; KL(q||v)=0.021) 5 months ago
.gitattributes

49.4 kB
Back up retrained Gemma-2-2B sample_rand_C1 (no final soft-cap, phase2 step 400) 1 day ago
README.md

1.32 kB
Upload README.md with huggingface_hub 5 months ago