Why a high kappa score still does not make an LLM judge trustworthy
IanArawjo · x · 2026-07-21
A follow-up explanation of the same metric issue: agreement statistics are more sensitive to random disagreements than to a consistent offset.
If an LLM is biased by a nearly constant amount — for example, always scoring one point above or below human raters — inter-rater alignment metrics may still look fine. That means a high kappa score does not necessarily mean the judge is truly reliable.
Takeaway
- Random noise hurts agreement scores more than stable bias.
- A judge can look “aligned” while still being systematically off in practice.
Related event: Study Warns: Unchecked LLM Judges Yield High False Positives(7 posts)→
More from Research
- MaP-WAM tackles non-Markovian robot manipulation with memory-grounded planning — Sizhe Zhao · 2026-09-11
- Negative Self-Distillation improves LLM reasoning by avoiding flawed reasoning paths — Rongcan Pei · 2026-09-11
- DeepMind-led paper makes design docs the source of truth, code disposable — SMART regenerates in 1.5-3h for ~$100 — Roger_M_Taylor · 2026-09-11
- GameWorld wins Best Paper Runner-Up at ECCV 2026 Multimodal Digital Agents Workshop — MikeShou1 · 2026-09-11
- Yann LeCun live at ECCV on World Models — Weak_Assistance_5261 · 2026-09-11
- 3D ResNet Paper Crosses 3,000 Citations Eight Years After CVPR 2018 — HirokatuKataoka · 2026-09-11