TI_var Predicts Judge Alignment
sanmikoyejo · x · 2026-07-12
The thread shares TIvar's predictive results for judge-level human alignment: on the validation set, the Spearman correlation is 0.833 for average labels, 0.905 for pass1, and 0.738 for pass2. Experiments cover judge families including GPT-5.3-Codex, Claude Opus 4.7, Qwen3-Coder.
More from Research
- Proto-tools, a universal infrastructure layer for generative biology, now open source — jmschreiber91 · 2026-10-26
- New paper finds AI agents negotiate badly under uncertainty and break deals on purpose — xuanalogue · 2026-10-26
- MCP fault injection study: agent task scores drop 60% under realistic tool failures, scale doesn't help — dair_ai · 2026-10-26
- Study: AI reviewing AI misses most reward hacking; best models miss 25%+ of hacks — ronbodkin · 2026-10-26
- FLUX.3 Klein step-distilled to 4-8 steps with CFG baked in — Fit-Structure-6489 · 2026-10-26
- NLI alone rejects 5/6 new agent memories; a 0.45 similarity gate fixes it all — Alternative_Pin9598 · 2026-10-26