Towards a Holistic and Automated Evaluation Framework for Multi-Level Comprehension of LLMs in Book-Length Contexts
This work addresses the challenge of automated evaluation for multi-level comprehension in LLMs, which is incremental as it builds on existing evaluation methods by introducing a structured, hierarchical approach.
The authors tackled the problem of evaluating long-context comprehension in LLMs by introducing HAMLET, a framework that structures texts into a three-level hierarchy and uses query-focused summarization, revealing that LLMs struggle with fine-grained comprehension and are sensitive to positional effects, with automatic evaluation achieving over 90% agreement with human judgments and reducing costs by up to 25 times.
We introduce HAMLET, a holistic and automated framework for evaluating the long-context comprehension of large language models (LLMs). HAMLET structures source texts into a three-level key-fact hierarchy at root-, branch-, and leaf-levels, and employs query-focused summarization to evaluate how well models recall and faithfully represent information at each level. To validate the reliability of our fully automated pipeline, we conduct a systematic human study, showing that our automatic evaluation achieves over 90% agreement with expert human judgments, while reducing the cost by up to 25 times. HAMLET reveals that LLMs struggle with fine-grained comprehension, especially at the leaf level, and are sensitive to positional effects like the lost-in-the-middle. Analytical queries pose greater challenges than narrative ones, and consistent performance gaps emerge between open-source and proprietary models, as well as across model scales. Our code and dataset are publicly available at https://github.com/DISL-Lab/HAMLET.