Capturing Failures of Large Language Models via Human Cognitive Biases

Erik Jones, Jacob Steinhardt

Keywords: open-ended generation, ai safety, codex, robustness, cognitive biases

Abstract: Large language models generate complex, open-ended outputs: instead of outputting a class label they write summaries, generate dialogue, or produce working code. In order to asses the reliability of these open-ended generation systems, we aim to identify qualitative categories of erroneous behavior, beyond identifying individual errors. To hypothesize and test for such qualitative errors, we draw inspiration from human cognitive biases---systematic patterns of deviation from rational judgement. Specifically, we use cognitive biases as motivation to (i) generate hypotheses for problems that models may have, and (ii) develop experiments that elicit these problems. Using code generation as a case study, we find that OpenAI’s Codex errs predictably based on how the input prompt is framed, adjusts outputs towards anchors, and is biased towards outputs that mimic frequent training examples. We then use our framework to elicit high-impact errors such as incorrectly deleting files. Our results indicate that experimental methodology from cognitive science can help characterize how machine learning systems behave.

TL;DR: We identify qualitative failure modes of large code and language models by extending experimental methodology used to elicit cognitive biases.

Community Implementations: [ 1 code implementation](https://www.catalyzex.com/paper/capturing-failures-of-large-language-models/code)