Views
No views yet
meta-llama/Llama-3.1-8B-Instruct, trained with the
bidirectional anchor reward (p(y|x) anchor coefficient
0.1 + gpt-oss-120b verifier fallback bonus 0.1 (threshold 5)).1from transformers import AutoModelForCausalLM, AutoTokenizer
2model = AutoModelForCausalLM.from_pretrained("just1nseo/llama31-tulu3-8b-dpo-if-rlvr-anchor-judge", subfolder="global_step_91")
3tok = AutoTokenizer.from_pretrained("just1nseo/llama31-tulu3-8b-dpo-if-rlvr-anchor-judge", subfolder="global_step_91")| subfolder | epoch | uploaded (UTC) |
|---|---|---|
global_step_91 | 1 | 2026-08-10 02:10:21 |
global_step_182 | 2 | 2026-08-10 16:16:45 |
global_step_273 | 3 | 2026-08-11 06:40:08 |
global_step_364 | 4 | 2026-08-11 20:49:32 |
| experiment | llama31_tulu3_8b_dpo_grpo_nonthink_anchor_pyx01_llmverifier_gptoss120b_fallback_bonus01_threshold5_b1024_c1_t1_2k |
| anchor cache | if_ref_anchor_teacher_llama31_tulu3_8b_dpo_nonreason_train_seed1_val512_t1_p095_r2048_scored_by_llama31_tulu3_8b_dpo.json (v3, 93993 rows all complete) |
| steps / epoch | 91 |
| train batch size | 1024 |
| rollout n | 8 |
| max prompt / response | 2048 / 2048 |
| wandb | ifif/verl_if_rlvr |