Qwen3-8B-ODA-Mixture-100k
qwen3_1.7b_new_sudoku_one_action_C_sft_lr_5e_6__step_5004
qwen3_1.7b_new_sudoku_one_action_C_sft_lr_5e_6__step_3336
Affine-GTRbeatEVERYTHING
gemma-2-2b-it-cpt-fft
qwen3-4b-thinking-aimo-numina-cot-sft
final-d2-1.7b
merge_cosfmt_MRL4096_ROLLOUT4_LR5e-7_w0.5_pcb
merge_lenfmt_MRL4096_ROLLOUT4_LR5e-7_w0.5_pcb
merge_accfmt_MRL4096_ROLLOUT4_LR1e-6_w0.5_pcb
merge_accfmt_MRL4096_ROLLOUT4_LR5e-7_w0.5_tall_mask_ta
merge_cosfmt_MRL4096_ROLLOUT4_LR5e-7_w0.5_tall_mask_ta
merge_lenfmt_MRL4096_ROLLOUT4_LR5e-7_w0.5_tall_mask_ta
Thinkanywhere-mini-swe-agent
Quelix-8B-v0.1
dr-tulu-shortform-rl-400step
Affine-Very-5EZeKjmJRgsyf5AuozJUNrgdC7WB3BynzCCxbbcMyHXQvHdu
Qwen3-4B-Thinking-2507-GRPO-exp03
qwen3_1.7b_new_sudoku_one_action_C_sft_lr_5e_6__step_1668
lora_model_qwen3_kaggle_2_epoch
qwen-hanabi-merged
final-01-03
tieto-code-mini-4b-instruct
soul-agent
qwen3_1.7b_new_sudoku_one_action_C_sft_lr_5e_6__step_6672
Llama3B-Cot
affine-yaz125-5HYt2PcdrvNCKw3ndgzMNBhh7znMj6P4jKGzhmfwiwN63y7h
affine-aaa
qwq_mixed_evol8k_aug4k_1e5
qwen3_1.7b_sudoku_multi_action_easy_11_20_epoch1
SkeptiSTEM-4B-v2-R123-fully-merged-16bit
llama_rand_30pct
arc-abs-sft-no-oracle-lr5e-6-ep1-0104
qwen3_32B_embrace_cpt_IV_e1_synthetic_context_2_merged_16bit
short_paper_qwent_0.json_train_grpo_v3_dev
llama3_1_8b_dpo-1k_ED
short_paper_qwent_qwen3-thinking-4b_train_sft_all_train_no_think
short_paper_qwen_0.json_train_dpo_v1_dev
4b_SFT_NEW
Affine-first
Affine-top_v4
affine-update-27
