0

Instructed to Bias: Instruction-Tuned Language Models Exhibit Emergent Cognitive Bias

Instruction-tuned large language models exhibit cognitive biases, such as the decoy effect, certainty effect, and belief bias, absent or less pronounced in their pretrained counterparts.

Year
2023
Venue
arXiv 2023
Authors
4
Hosting
Abstract onlyARXIV-DEFAULT

Cite

Notes

Only stored in your browser.

Attribution

Abstract & full text
arxiv.org/abs/2308.00225v2ARXIV-DEFAULT
TL;DR
Semantic Scholar
Attribution policy →

Abstract

Recent studies show that instruction tuning (IT) and reinforcement learning from human feedback (RLHF) improve the abilities of large language models (LMs) dramatically. While these tuning methods can help align models with human objectives and generate high-quality text, not much is known about their potential adverse effects. In this work, we investigate the effect of IT and RLHF on decision making and reasoning in LMs, focusing on three cognitive biases - the decoy effect, the certainty effect, and the belief bias - all of which are known to influence human decision-making and reasoning. Our findings highlight the presence of these biases in various models from the GPT-3, Mistral, and T5 families. Notably, we find a stronger presence of biases in models that have undergone instruction tuning, such as Flan-T5, Mistral-Instruct, GPT3.5, and GPT4. Our work constitutes a step toward comprehending cognitive biases in instruction-tuned LMs, which is crucial for the development of more reliable and unbiased language models.

Authors

4