fine-grained evaluation
Fine-grained evaluation refers to assessing model performance at a detailed level, such as by examining specific classes, errors, or metrics, to gain deeper insights into model behavior and areas for improvement.
- AGI-Elo: How Far Are We From Mastering A Task?
- Are Large Reasoning Models Good Translation Evaluators? Analysis and Performance Boost
- ExAct: A Video-Language Benchmark for Expert Action Analysis
- KRIS-Bench: Benchmarking Next-Level Intelligent Image Editing Models
- OneIG-Bench: Omni-dimensional Nuanced Evaluation for Image Generation
- Quantifying Generalisation in Imitation Learning