Benchmark

F1-R²

F1-R² is an evaluation metric designed for the TestHallVQA benchmark. It jointly quantifies a large vision-language model's computational reasoning capability and its robustness in evidence retrieval under document-level redundancy, providing finer-grained model diagnostics than the traditional F1 score.

Timeline (last 90 days)

Sep 15

The research team proposed a new evaluation metric, F1-R², for jointly quantifying computational reasoning ability and evidence retrieval robustness under document-level redundancy.

Unverified50%

All Facts (1)

Unverified

The research team proposed a new evaluation metric, F1-R², for jointly quantifying computational reasoning ability and evidence retrieval robustness under document-level redundancy.

50%

Source Articles

TestHallVQA: A New Document-Level Reasoning Benchmark for Large Vision-Language Models Under Redundant ContextSep 15