Modelpublic

bowen0815/Compass-Failed-Attempts

sourceHugging Faceupdated 1mo agoView on Hugging Face

0likes

dir2026-04-29_math_process_feedback_dpo_pairs_v0/

dir2026-04-29_math_process_feedback_dpo_train_v0/

dir2026-04-29_same_prefix_dpo_pairs_v0/

dir2026-04-29_same_prefix_dpo_pairs_v1/

dir2026-04-29_same_prefix_dpo_train_v0/

dir2026-05-02_gpt54_v1_action_only_balanced_controller/

dir2026-05-03_gpt54_v7_controller_v2_stop_continue_review_balanced_gpu4/

dir2026-05-05_branch118_stop180_control50_branchrepeat2_gpu3/

dir2026-05-05_gpt55_branch_decision_all134_sftinit_dpo_gpu3/

dir2026-05-05_gpt55_branch_decision_all134_thought_gpu3/

dir2026-05-05_gpt55_branch_decision_pilot30_gpu3/

dir2026-05-05_gpt55_branch_decision_pilot30_thought_gpu3/

dir2026-05-05_gpt55_branch118_next3_sftinit_dpo_gpu3/

dir2026-05-05_gpt55_rank_basegen_all118_k5_256_gpu0/

dir2026-05-05_gpt55_rank_basegen_all118_k5_256_sftinit_dpo_gpu0/

dir2026-05-06_gpt55_branch133_stop250_control117_gpu0/

dir2026-05-07_gpt55_branch133_evidence_strategy_local_unboxed_gpu1/

dir2026-05-07_gpt55_branch133_next4init_onpolicy_dpo_trainonly_gpu1/

dir2026-05-08_gpt55_branch133_evidence_strategy_local_unboxed_gpu1/

dir2026-05-08_next3_stage1_hidden_evidence_w005_step80_gpu1_v2/

dir2026-05-08_smoke_candidates_train2_k2_records/

dir2026-05-08_smoke_student_dpo_step1/

dir2026-05-08_smoke_student_pref_pairs_train2_k2/

dir2026-05-08_stage1_hidden_evidence_smoke1_gpu1/

dir2026-05-19_qwen7500_wrong_remaining_candidates/

dirmath_process_feedback_clean_v1/

dirmath_process_feedback_clean_v1_full_current/

dirmath_process_feedback_clean_v1_per80_current/

dirmath_process_feedback_clean_v2_8b_all_current/

dirpre_resume_step3_backup_20260601_153723/

dirqwen3_4b_gpt_union_control85_plus_hard20_local_continuation_drop_stop_lb_sft220_lb6/

dirqwen3_4b_gpt_union_control85_plus_hard20_local_continuation_sft220_lb6/

dirqwen3_4b_gsm8k_control_pilot101_action_only_sft80_lb10_stop2/

dirqwen3_4b_gsm8k_control_pilot101_anchor_only_sft120/

dirqwen3_4b_gsm8k_control_pilot101_anchor_only_sft40_control2_stop8/

dirqwen3_4b_gsm8k_flashtrace_evidence_control85_sft80_lb10_stop2/

dirqwen3_4b_gsm8k_flashtrace_gpt_union_ids_control85_plus_hard20_sft120_lb10_stop2/

dirqwen3_4b_gsm8k_flashtrace_gpt_union_ids_control85_plus_hard20_sft160_lb4_stop1/

dirqwen3_4b_gsm8k_flashtrace_ids_control85_plus_hard20_sft120_lb10_stop2/

dirqwen3_4b_gsm8k_flashtrace_ids_control85_sft80_lb10_stop2/

file.gitattributes5 KBdownload

file2026-05-08_next3_evidence_causal_smoke2_generated2.json10 KBdownload

file2026-05-08_smoke_candidates_train2_k2.json14 KBdownload

file2026-05-08_smoke_gates_score_only.json547 Bdownload

file2026-05-08_smoke_gpt_rank_train2_k2.json1 KBdownload

file2026-05-08_smoke_scores_train2_k2_flashtrace.json28 KBdownload

file2026-05-09_gpt55_branch133_localverify_next4_oldinit_sft40_lr1e6_replay03_gpu1.train.log15 KBdownload

file2026-05-09_gpt55_branch133_next3_stop_eos_oldinit_sft80_lr2e6_gpu1.train.log26 KBdownload

file2026-05-09_gpt55_branch133_next3_stoprows_oldinit_sft80_lr2e6_stop8_gpu1.train.log26 KBdownload

file2026-05-20_direct_branch_clue_reconstruct_train211.log30 KBdownload

file2026-05-20_G_sft_train211.log21 KBdownload

filemetrics.jsonl7 KBdownload

fileoffline_eval_plan.json476 Bdownload

fileprompt_length_distribution.json686 Bdownload

fileresolved_config.yaml28 KBdownload

filerun.log125 KBdownload

bowen0815/Compass-Failed-Attempts · main · files are served by the source, never re-hosted here