#evaluation
Evaluation: 4 AI articles covering evaluation news, analysis, and research
Articles
ParseBench-2: A Severity-Aware Benchmark for Document Parsingβ8
ParseBench-2 introduces a severity-aware benchmark for document parsing, with 4,317 pages and metrics calibrated to error severity across text, tables, charts, ...
How UK AISI and EvalEval Are Making AI Benchmark Results Reproducibleβ9
UK AISI and EvalEval are building standards and infrastructure to make AI benchmark results reproducible, traceable, and comparable across institutions.
What Do We Expect from LLMs? A Systematic Map of LLM Benchmarkβ9
This systematic map of 14,767 LLM benchmark papers (2022β2026) reveals shifting capability expectations, rising model participation, and risks of evaluator bias...
Position: Reasoning is a Learnable Rule-Based Processβ10
Reasoning in AI needs clear definitions. This paper positions it as a learnable rule-based process and proposes best practices for evaluating and communicating ...
