Saved in:
Bibliographic Details
Main Authors: Du, Kevin, Snæbjarnarson, Vésteinn, Stoehr, Niklas, White, Jennifer C., Schein, Aaron, Cotterell, Ryan
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2404.04633
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909224428109824
author Du, Kevin
Snæbjarnarson, Vésteinn
Stoehr, Niklas
White, Jennifer C.
Schein, Aaron
Cotterell, Ryan
author_facet Du, Kevin
Snæbjarnarson, Vésteinn
Stoehr, Niklas
White, Jennifer C.
Schein, Aaron
Cotterell, Ryan
contents To answer a question, language models often need to integrate prior knowledge learned during pretraining and new information presented in context. We hypothesize that models perform this integration in a predictable way across different questions and contexts: models will rely more on prior knowledge for questions about entities (e.g., persons, places, etc.) that they are more familiar with due to higher exposure in the training corpus, and be more easily persuaded by some contexts than others. To formalize this problem, we propose two mutual information-based metrics to measure a model's dependency on a context and on its prior about an entity: first, the persuasion score of a given context represents how much a model depends on the context in its decision, and second, the susceptibility score of a given entity represents how much the model can be swayed away from its original answer distribution about an entity. We empirically test our metrics for their validity and reliability. Finally, we explore and find a relationship between the scores and the model's expected familiarity with an entity, and provide two use cases to illustrate their benefits.
format Preprint
id arxiv_https___arxiv_org_abs_2404_04633
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Context versus Prior Knowledge in Language Models
Du, Kevin
Snæbjarnarson, Vésteinn
Stoehr, Niklas
White, Jennifer C.
Schein, Aaron
Cotterell, Ryan
Computation and Language
To answer a question, language models often need to integrate prior knowledge learned during pretraining and new information presented in context. We hypothesize that models perform this integration in a predictable way across different questions and contexts: models will rely more on prior knowledge for questions about entities (e.g., persons, places, etc.) that they are more familiar with due to higher exposure in the training corpus, and be more easily persuaded by some contexts than others. To formalize this problem, we propose two mutual information-based metrics to measure a model's dependency on a context and on its prior about an entity: first, the persuasion score of a given context represents how much a model depends on the context in its decision, and second, the susceptibility score of a given entity represents how much the model can be swayed away from its original answer distribution about an entity. We empirically test our metrics for their validity and reliability. Finally, we explore and find a relationship between the scores and the model's expected familiarity with an entity, and provide two use cases to illustrate their benefits.
title Context versus Prior Knowledge in Language Models
topic Computation and Language
url https://arxiv.org/abs/2404.04633