A Peek into Token Bias: Large Language Models Are Not Yet Genuine Reasoners
This work addresses concerns about the actual reasoning and generalization capabilities of LLMs, which is critical for AI researchers and developers, though it is incremental in testing existing models.
The study investigated whether large language models (LLMs) possess genuine reasoning abilities or rely on token bias, finding that most LLMs struggle with logical reasoning and depend on superficial patterns with strong token bias.
This study introduces a hypothesis-testing framework to assess whether large language models (LLMs) possess genuine reasoning abilities or primarily depend on token bias. We go beyond evaluating LLMs on accuracy; rather, we aim to investigate their token bias in solving logical reasoning tasks. Specifically, we develop carefully controlled synthetic datasets, featuring conjunction fallacy and syllogistic problems. Our framework outlines a list of hypotheses where token biases are readily identifiable, with all null hypotheses assuming genuine reasoning capabilities of LLMs. The findings in this study suggest, with statistical guarantee, that most LLMs still struggle with logical reasoning. While they may perform well on classic problems, their success largely depends on recognizing superficial patterns with strong token bias, thereby raising concerns about their actual reasoning and generalization abilities. Codes and data are open-sourced at https://github.com/bowen-upenn/llm_token_bias.