LLM-as-a-Coach turns judge feedback into transferable experiential knowledge
iScienceLuvr · x · 2026-07-21
The paper proposes Experiential Learning (EL), which turns the usual LLM-as-a-Judge setup into an LLM-as-a-Coach.
Instead of collapsing feedback into a scalar score, the coach distills its assessment of each on-policy response into transferable experiential knowledge. That knowledge conditions a teacher model and is then internalized by the policy through on-policy context distillation. The authors report that, compared with standard rubric-based RL, EL improves held-out and unseen open-ended tasks, generalizes better beyond the training distribution, and helps mitigate reward hacking.
The screenshot also shows the paper’s high-level claim: richer, higher-bandwidth feedback preserves fine-grained preferences among high-quality responses and works across two policy families.
Related event: Microsoft Proposes LLM-as-a-Coach for Non-Verifiable Tasks(4 posts)→
More from Research
- Diamandis: AI Cancer Navigator and $100 Genome Sequencing Close the Care Gap — PeterDiamandis · 2026-10-27
- EMNLP Oral: training-free Nexus Sampling cuts KV-cache memory 10x; OpenResearcher tops BrowseComp-Plus at 54.8% — yuz9yuz · 2026-10-27
- ICML 2026 position paper: LLM 'thinking traces' are a harmful anthropomorphism — rao2z · 2026-10-27
- Forking gap follows a power law in context frequency; masking rare n-grams shrinks it — ZimingLiu11 · 2026-10-27
- How forking happens: n-gram tables split logits into non-interacting subspaces — ZimingLiu11 · 2026-10-27
- The best LoRA explainer video yet: rank constraints meet intrinsic dimension — sophiamyang · 2026-10-27