List your company

User-aligned task completion on KitchenAmbig (In-Distribution)

84AccuracyAdaptive CLIPR
60.49666.59872.778.802May 12, 2026

Evaluation results

MethodLinks
Adaptive CLIPRLLM engine=Claude Sonnet 4.5, alpha=1.5, f=52026.0584
GATELLM engine=Claude Sonnet 4.52026.0569.5
CIPHERLLM engine=Claude Sonnet 4.52026.0561.4