Towards a Theoretical Understanding of the 'Reversal Curse' via Training Dynamics

arXiv:2405.04669v1 Announce Type: new
Abstract: Auto-regressive large language models (LLMs) show impressive capacities to solve many complex reasoning tasks while struggling with some simple logical reasoning tasks such as inverse search: when trained on ''A is B'', LLM fails to directly conclude ''B is A'' during inference, which is known as the ''reversal curse'' (Berglund et al., 2023). In this paper, we theoretically analyze the reversal curse via the training dynamics of (stochastic) gradient descent for two auto-regressive models: (1) …

abstract arxiv auto cs.lg dynamics inference language language models large language large language models llm llms reasoning search show simple solve tasks training type understanding via while