Your judge is drifting too — why evaluation standards move even when the model doesn't
This digest covers LLM evaluation and observability developments from 2026-11-05 through 2026-11-12. The headline thread this week: several independent papers converge on the same uncomfortable finding — the instruments we use to measure LLM quality (human raters, LLM-judges, contamination tests, f…
Read the full story at DEV Community — Machine Learning ↗
Timeline · 1 report
- 2026-11-12 12:14 · DEV Community — Machine Learning
Your judge is drifting too — why evaluation standards move even when the model doesn't