A Closer Look at Claim Decomposition
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911801044631552 |
|---|---|
| author | Wanner, Miriam Ebner, Seth Jiang, Zhengping Dredze, Mark Van Durme, Benjamin |
| author_facet | Wanner, Miriam Ebner, Seth Jiang, Zhengping Dredze, Mark Van Durme, Benjamin |
| contents | As generated text becomes more commonplace, it is increasingly important to evaluate how well-supported such text is by external knowledge sources. Many approaches for evaluating textual support rely on some method for decomposing text into its individual subclaims which are scored against a trusted reference. We investigate how various methods of claim decomposition -- especially LLM-based methods -- affect the result of an evaluation approach such as the recently proposed FActScore, finding that it is sensitive to the decomposition method used. This sensitivity arises because such metrics attribute overall textual support to the model that generated the text even though error can also come from the metric's decomposition step. To measure decomposition quality, we introduce an adaptation of FActScore, which we call DecompScore. We then propose an LLM-based approach to generating decompositions inspired by Bertrand Russell's theory of logical atomism and neo-Davidsonian semantics and demonstrate its improved decomposition quality over previous methods. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_11903 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | A Closer Look at Claim Decomposition Wanner, Miriam Ebner, Seth Jiang, Zhengping Dredze, Mark Van Durme, Benjamin Computation and Language As generated text becomes more commonplace, it is increasingly important to evaluate how well-supported such text is by external knowledge sources. Many approaches for evaluating textual support rely on some method for decomposing text into its individual subclaims which are scored against a trusted reference. We investigate how various methods of claim decomposition -- especially LLM-based methods -- affect the result of an evaluation approach such as the recently proposed FActScore, finding that it is sensitive to the decomposition method used. This sensitivity arises because such metrics attribute overall textual support to the model that generated the text even though error can also come from the metric's decomposition step. To measure decomposition quality, we introduce an adaptation of FActScore, which we call DecompScore. We then propose an LLM-based approach to generating decompositions inspired by Bertrand Russell's theory of logical atomism and neo-Davidsonian semantics and demonstrate its improved decomposition quality over previous methods. |
| title | A Closer Look at Claim Decomposition |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2403.11903 |