F-Eval: Assessing Fundamental Abilities with Refined Evaluation Methods
This work addresses the need for better evaluation of fundamental abilities in LLMs, which is incremental as it refines existing evaluation methods rather than introducing a new paradigm.
The paper tackles the problem that existing evaluation benchmarks for large language models (LLMs) focus on instruction-following capabilities and overlook fundamental abilities from pre-training, proposing F-Eval, a bilingual benchmark to assess expression, commonsense, and logic with new evaluation methods for reference-free subjective tasks, resulting in higher correlation coefficients and larger distinction than other evaluators on 13 advanced LLMs.
Large language models (LLMs) garner significant attention for their unprecedented performance, leading to an increasing number of researches evaluating LLMs. However, these evaluation benchmarks are limited to assessing the instruction-following capabilities, overlooking the fundamental abilities that emerge during the pre-training stage. Previous subjective evaluation methods mainly reply on scoring by API models. However, in the absence of references, large models have shown limited ability to discern subtle differences. To bridge the gap, we propose F-Eval, a bilingual evaluation benchmark to evaluate the fundamental abilities, including expression, commonsense and logic. The tasks in F-Eval include multi-choice objective tasks, open-ended objective tasks, reference-based subjective tasks and reference-free subjective tasks. For reference-free subjective tasks, we devise new evaluation methods, serving as alternatives to scoring by API models. We conduct evaluations on 13 advanced LLMs. Results show that our evaluation methods show higher correlation coefficients and larger distinction than other evaluators. Additionally, we discuss the influence of different model sizes, dimensions, and normalization methods. We anticipate that F-Eval will facilitate the study of LLMs' fundamental abilities.