Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2406.06369 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866913535259312128 |
|---|---|
| author | Movva, Rajiv Koh, Pang Wei Pierson, Emma |
| author_facet | Movva, Rajiv Koh, Pang Wei Pierson, Emma |
| contents | Do LLMs align with human perceptions of safety? We study this question via annotation alignment, the extent to which LLMs and humans agree when annotating the safety of user-chatbot conversations. We leverage the recent DICES dataset (Aroyo et al., 2023), in which 350 conversations are each rated for safety by 112 annotators spanning 10 race-gender groups. GPT-4 achieves a Pearson correlation of $r = 0.59$ with the average annotator rating, \textit{higher} than the median annotator's correlation with the average ($r=0.51$). We show that larger datasets are needed to resolve whether LLMs exhibit disparities in how well they correlate with different demographic groups. Also, there is substantial idiosyncratic variation in correlation within groups, suggesting that race & gender do not fully capture differences in alignment. Finally, we find that GPT-4 cannot predict when one demographic group finds a conversation more unsafe than another. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_06369 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Annotation alignment: Comparing LLM and human annotations of conversational safety Movva, Rajiv Koh, Pang Wei Pierson, Emma Computation and Language Do LLMs align with human perceptions of safety? We study this question via annotation alignment, the extent to which LLMs and humans agree when annotating the safety of user-chatbot conversations. We leverage the recent DICES dataset (Aroyo et al., 2023), in which 350 conversations are each rated for safety by 112 annotators spanning 10 race-gender groups. GPT-4 achieves a Pearson correlation of $r = 0.59$ with the average annotator rating, \textit{higher} than the median annotator's correlation with the average ($r=0.51$). We show that larger datasets are needed to resolve whether LLMs exhibit disparities in how well they correlate with different demographic groups. Also, there is substantial idiosyncratic variation in correlation within groups, suggesting that race & gender do not fully capture differences in alignment. Finally, we find that GPT-4 cannot predict when one demographic group finds a conversation more unsafe than another. |
| title | Annotation alignment: Comparing LLM and human annotations of conversational safety |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2406.06369 |