Counterfactual Probing for the Influence of Affect and Specificity on Intergroup Bias
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866917412145725440 |
|---|---|
| author | Govindarajan, Venkata S Mahowald, Kyle Beaver, David I. Li, Junyi Jessy |
| author_facet | Govindarajan, Venkata S Mahowald, Kyle Beaver, David I. Li, Junyi Jessy |
| contents | While existing work on studying bias in NLP focues on negative or pejorative language use, Govindarajan et al. (2023) offer a revised framing of bias in terms of intergroup social context, and its effects on language behavior. In this paper, we investigate if two pragmatic features (specificity and affect) systematically vary in different intergroup contexts -- thus connecting this new framing of bias to language output. Preliminary analysis finds modest correlations between specificity and affect of tweets with supervised intergroup relationship (IGR) labels. Counterfactual probing further reveals that while neural models finetuned for predicting IGR labels reliably use affect in classification, the model's usage of specificity is inconclusive. Code and data can be found at: https://github.com/venkatasg/intergroup-probing |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2305_16409 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Counterfactual Probing for the Influence of Affect and Specificity on Intergroup Bias Govindarajan, Venkata S Mahowald, Kyle Beaver, David I. Li, Junyi Jessy Computation and Language Computers and Society While existing work on studying bias in NLP focues on negative or pejorative language use, Govindarajan et al. (2023) offer a revised framing of bias in terms of intergroup social context, and its effects on language behavior. In this paper, we investigate if two pragmatic features (specificity and affect) systematically vary in different intergroup contexts -- thus connecting this new framing of bias to language output. Preliminary analysis finds modest correlations between specificity and affect of tweets with supervised intergroup relationship (IGR) labels. Counterfactual probing further reveals that while neural models finetuned for predicting IGR labels reliably use affect in classification, the model's usage of specificity is inconclusive. Code and data can be found at: https://github.com/venkatasg/intergroup-probing |
| title | Counterfactual Probing for the Influence of Affect and Specificity on Intergroup Bias |
| topic | Computation and Language Computers and Society |
| url | https://arxiv.org/abs/2305.16409 |