New Thematic Generalization Benchmark: measures how effectively LLMs infer a specific "theme" from a small set of examples and anti-examples

/u/zero0_one1

January 14, 2025 January 14, 2025