A Practical Method for Generating String Counterfactuals
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866913685453144064 |
|---|---|
| author | Avitan, Matan Cotterell, Ryan Goldberg, Yoav Ravfogel, Shauli |
| author_facet | Avitan, Matan Cotterell, Ryan Goldberg, Yoav Ravfogel, Shauli |
| contents | Interventions targeting the representation space of language models (LMs) have emerged as an effective means to influence model behavior. Such methods are employed, for example, to eliminate or alter the encoding of demographic information such as gender within the model's representations and, in so doing, create a counterfactual representation. However, because the intervention operates within the representation space, understanding precisely what aspects of the text it modifies poses a challenge. In this paper, we give a method to convert representation counterfactuals into string counterfactuals. We demonstrate that this approach enables us to analyze the linguistic alterations corresponding to a given representation space intervention and to interpret the features utilized to encode a specific concept. Moreover, the resulting counterfactuals can be used to mitigate bias in classification through data augmentation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_11355 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | A Practical Method for Generating String Counterfactuals Avitan, Matan Cotterell, Ryan Goldberg, Yoav Ravfogel, Shauli Computation and Language Computers and Society Machine Learning Interventions targeting the representation space of language models (LMs) have emerged as an effective means to influence model behavior. Such methods are employed, for example, to eliminate or alter the encoding of demographic information such as gender within the model's representations and, in so doing, create a counterfactual representation. However, because the intervention operates within the representation space, understanding precisely what aspects of the text it modifies poses a challenge. In this paper, we give a method to convert representation counterfactuals into string counterfactuals. We demonstrate that this approach enables us to analyze the linguistic alterations corresponding to a given representation space intervention and to interpret the features utilized to encode a specific concept. Moreover, the resulting counterfactuals can be used to mitigate bias in classification through data augmentation. |
| title | A Practical Method for Generating String Counterfactuals |
| topic | Computation and Language Computers and Society Machine Learning |
| url | https://arxiv.org/abs/2402.11355 |