Language models are not naysayers: An analysis of language models on negation benchmarks
This addresses a fundamental linguistic problem for NLP researchers and practitioners, but it is incremental as it extends prior work on negation from masked models to auto-regressive LLMs.
The paper evaluated large language models (LLMs) like GPT-neo, GPT-3, and InstructGPT on negation benchmarks, finding that they have limitations such as insensitivity to negation, inability to capture its lexical semantics, and failure to reason under negation.
Negation has been shown to be a major bottleneck for masked language models, such as BERT. However, whether this finding still holds for larger-sized auto-regressive language models (``LLMs'') has not been studied comprehensively. With the ever-increasing volume of research and applications of LLMs, we take a step back to evaluate the ability of current-generation LLMs to handle negation, a fundamental linguistic phenomenon that is central to language understanding. We evaluate different LLMs -- including the open-source GPT-neo, GPT-3, and InstructGPT -- against a wide range of negation benchmarks. Through systematic experimentation with varying model sizes and prompts, we show that LLMs have several limitations including insensitivity to the presence of negation, an inability to capture the lexical semantics of negation, and a failure to reason under negation.