Lost in Inference: Rediscovering the Role of Natural Language Inference for Large Language Models

TL;DR

This paper explores the effectiveness of natural language inference (NLI) tasks in evaluating large language models, showing they can discriminate model quality and align with human judgments better as models scale.

Contribution

It demonstrates that NLI benchmarks are valuable for assessing LLMs, especially in distinguishing different model sizes and understanding their training progress.

Findings

NLI tasks effectively differentiate models of various sizes.

Model-human distribution similarity increases with model scale.

NLI benchmarks are not fully saturated and remain informative.

Abstract

In the recent past, a popular way of evaluating natural language understanding (NLU), was to consider a model's ability to perform natural language inference (NLI) tasks. In this paper, we investigate if NLI tasks, that are rarely used for LLM evaluation, can still be informative for evaluating LLMs. Focusing on five different NLI benchmarks across six models of different scales, we investigate if they are able to discriminate models of different size and quality and how their accuracies develop during training. Furthermore, we investigate the extent to which the softmax distributions of models align with human distributions in cases where statements are ambiguous or vague. Overall, our results paint a positive picture for the NLI tasks: we find that they are able to discriminate well between models at various stages of training, yet are not (all) saturated. Furthermore, we find that while the similarity of model distributions with human label distributions increases with scale, it is still much higher than the similarity between two populations of humans, making it a potentially interesting statistic to consider.