lechmazur/generalization

Thematic Generalization Benchmark: measures how effectively various LLMs can infer a narrow or specific "theme" (category/rule) from a small set of examples and anti-examples, then detect which item truly fits that theme among a collection of misleading candidates.

6 best points · 0 comments · 1 Hacker News post · 74 GitHub stars · View on GitHub

Hacker News discussions

• Show HN: LLM Thematic Generalization Benchmark (6 points, 0 comments, by zone411, 2025-01-14)