CoRT: Counterfactual Replay for Token-Level Rubric-Guided Policy Optimization
Paper • 2607.25659 • Published • 42 •
3
🔄 In a Training Loop
quinn
jwhe
AI & ML interests
None yet
Recent Activity
authored a paper 3 days ago
SMOPD: Multi-Reward Reinforcement Learning via Specialize-and-Merge Online Policy Distillation authored a paper 3 days ago
Harbor Adapters and Harbor-Index: Infrastructure and a Curated Meta-Dataset for Large-Scale Agentic Evaluation authored a paper 3 days ago
Scoring Higher, Answering Worse: Mitigating Reward Hacking in Rubric-Based RL via Protocol-Level RubricsOrganizations
commented a
paper 3 months ago
[Parity] CL-bench: codex/gpt-5.2 vs infer_codex.py (50 tasks, 3 trials, MATCHING)
1
#230 opened 6 months ago
by jwhe
[Parity] CL-bench: codex/gpt-5.1 vs original pipeline (50 tasks, 3 trials)
#210 opened 6 months ago
by jwhe