# Beyond Scores: Diagnostic LLM Evaluation via Fine-Grained Abilities

arXiv: https://arxiv.org/abs/2604.12191 · HTML: https://mlsanity.com/paper/2604.12191/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 6 of its 32 references that are in the map.
- First posted on arXiv 14 Apr 2026.

## Builds on (most cited first, 6 of 6)

- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [SciEval: A Multi-Level Large Language Model Evaluation Benchmark for Scientific Research](https://mlsanity.com/paper/2308.13149/) (2023)
- [Lost in Benchmarks? Rethinking Large Language Model Benchmarking with Item Response Theory](https://mlsanity.com/paper/2505.15055/) (2025)
- [Toward Generalizable Evaluation in the LLM Era: A Survey Beyond Benchmarks](https://mlsanity.com/paper/2504.18838/) (2025)
- [Can We Predict Performance of Large Models across Vision-Language Tasks?](https://mlsanity.com/paper/2410.10112/) (2024)
- [JE-IRT: A Geometric Lens on LLM Abilities through Joint Embedding Item Response Theory](https://mlsanity.com/paper/2509.22888/) (2025)

## Built on by (most cited first, 1 of 1)

- [Curriculum Brain: Constructing Curriculum Knowledge Graphs as a Substrate for Cognitive Diagnosis](https://mlsanity.com/paper/2610.05860/) (2026)

## Abstract (arXiv)

Current evaluations of large language models aggregate performance across diverse tasks into single scores. This obscures fine-grained ability variation, limiting targeted model improvement and ability-guided selection for specific tasks. Motivated by this gap, we propose a cognitive diagnostic framework that estimates model abilities across multiple fine-grained dimensions. For mathematics, we construct a 35-dimensional ability taxonomy grounded in cognitive theory and domain knowledge. The framework employs multidimensional Item Response Theory with an item-ability association matrix to estimate fine-grained ability levels, which in turn enable prediction of performance on unseen items (questions of benchmark). Evaluated on 41 models, our approach demonstrates strong criterion validity, consistent ability estimates across benchmarks, and accurate prediction of unseen items with AUC ranging from 0.80 to 0.89 within benchmarks and from 0.77 to 0.86 across benchmarks, substantially exceeding trivial baselines. The framework generalizes across scientific domains, producing consistent diagnostic performance in physics (27 dimensions), chemistry (58 dimensions), and computer science (12 dimensions). This work establishes a principled framework for fine-grained assessment of abilities, with potential applications in targeted training, ability-guided model selection, and ability-aware benchmark design.