From Understanding to Generation: An Efficient Shortcut for Evaluating Language Models
This work addresses the problem of efficient model evaluation for researchers and developers, enabling more frequent monitoring during training, though it is incremental as it adapts existing evaluation formats.
The paper tackles the high computational cost of evaluating language models on generative tasks by reformulating them into cheaper multiple-choice alternatives, achieving over 35x average reduction in evaluation time while maintaining strong performance correlation across various capabilities.
Iterative evaluation of LLMs during training is essential to ensure expected capability development, but can be time- and compute-intensive. While NLU tasks, where the model selects from fixed answer choices, are cheap to evaluate, essential capabilities like reasoning and code generation rely on the more time-consuming NLG (token-by-token generation) format. In this work, our aim is to decrease the computational burden of NLG benchmarks in order to enable monitoring crucial LLM capabilities during model training. We reformulate generative tasks into computationally cheaper NLU alternatives. We test the performance correlation between the original and reformulated tasks using 8 LMs of various sizes and 4 capabilities: mathematical reasoning, code generation, factual knowledge and reading comprehension. Our results show a strong correlation between task formats, supporting capability assessment via cheaper alternatives and achieving over 35x average reduction in evaluation time. Our project is available at: https://github.com/Fraunhofer-IIS/EvalShortcut