Rare Event Analysis of Large Language Models
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866918528583467008 |
|---|---|
| author | Dorman, Jake McAllister Gillman, Edward Rose, Dominic C. Mair, Jamie F. Garrahan, Juan P. |
| author_facet | Dorman, Jake McAllister Gillman, Edward Rose, Dominic C. Mair, Jamie F. Garrahan, Juan P. |
| contents | Being probabilistic models, during inference large language models (LLMs) display rare events: behaviour that is far from typical but highly significant. By definition all rare events are hard to see, but the enormous scale of LLM usage means that events completely unobserved during development are likely to become prominent in deployment. Here we present an end-to-end framework for the systematic analysis of rare events in LLMs. We provide a practical implementation spanning theory, efficient generation strategies, probability estimation and error analysis, which we illustrate with concrete examples. We outline extensions and applications to other models and contexts, highlighting the generality of the concepts and techniques presented here. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_06791 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Rare Event Analysis of Large Language Models Dorman, Jake McAllister Gillman, Edward Rose, Dominic C. Mair, Jamie F. Garrahan, Juan P. Machine Learning Disordered Systems and Neural Networks Statistical Mechanics Being probabilistic models, during inference large language models (LLMs) display rare events: behaviour that is far from typical but highly significant. By definition all rare events are hard to see, but the enormous scale of LLM usage means that events completely unobserved during development are likely to become prominent in deployment. Here we present an end-to-end framework for the systematic analysis of rare events in LLMs. We provide a practical implementation spanning theory, efficient generation strategies, probability estimation and error analysis, which we illustrate with concrete examples. We outline extensions and applications to other models and contexts, highlighting the generality of the concepts and techniques presented here. |
| title | Rare Event Analysis of Large Language Models |
| topic | Machine Learning Disordered Systems and Neural Networks Statistical Mechanics |
| url | https://arxiv.org/abs/2602.06791 |