Explicit Inductive Inference using Large Language Models
This work addresses a specific bias in LLMs for inference tasks, offering an incremental improvement in domain-specific applications.
The paper tackled the problem of LLMs' attestation bias in inference tasks by proposing a pipeline that transforms premises into attested alternatives to improve entailment predictions, achieving substantial performance gains on a directional predicate entailment benchmark.
Large Language Models (LLMs) are reported to hold undesirable attestation bias on inference tasks: when asked to predict if a premise P entails a hypothesis H, instead of considering H's conditional truthfulness entailed by P, LLMs tend to use the out-of-context truth label of H as a fragile proxy. In this paper, we propose a pipeline that exploits this bias to do explicit inductive inference. Our pipeline uses an LLM to transform a premise into a set of attested alternatives, and then aggregate answers of the derived new entailment inquiries to support the original inference prediction. On a directional predicate entailment benchmark, we demonstrate that by applying this simple pipeline, we can improve the overall performance of LLMs on inference and substantially alleviate the impact of their attestation bias.