AINov 16, 2023

Program-Aided Reasoners (better) Know What They Know

CMU
arXiv:2311.09553v134 citationsh-index: 21
Originality Incremental advance
AI Analysis

This work addresses the problem of ensuring large language models are well-calibrated for reasoning tasks, which is important for reliability in AI applications, though it is incremental as it builds on prior program-aided reasoning methods.

The paper compares the calibration of Program Aided Language Models (PAL) and text-based Chain-of-thought (COT) prompting across 5 datasets and 2 model types, finding that PAL leads to improved calibration in 75% of instances, with experiments showing that lower generation diversity enhances calibration.

Prior work shows that program-aided reasoning, in which large language models (LLMs) are combined with programs written in programming languages such as Python, can significantly improve accuracy on various reasoning tasks. However, while accuracy is essential, it is also important for such reasoners to "know what they know", which can be quantified through the calibration of the model. In this paper, we compare the calibration of Program Aided Language Models (PAL) and text-based Chain-of-thought (COT) prompting techniques over 5 datasets and 2 model types: LLaMA models and OpenAI models. Our results indicate that PAL leads to improved calibration in 75% of the instances. Our analysis uncovers that prompting styles that produce lesser diversity in generations also have more calibrated results, and thus we also experiment with inducing lower generation diversity using temperature scaling and find that for certain temperatures, PAL is not only more accurate but is also more calibrated than COT. Overall, we demonstrate that, in the majority of cases, program-aided reasoners better know what they know than text-based counterparts.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes