CoCoLex: Confidence-guided Copy-based Decoding for Grounded Legal Text Generation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911096139415552 |
|---|---|
| author | S, Santosh T. Y. S. Elkhayat, Youssef Tarek Ichim, Oana Shetty, Pranav Wang, Dongsheng Ma, Zhiqiang Nourbakhsh, Armineh Liu, Xiaomo |
| author_facet | S, Santosh T. Y. S. Elkhayat, Youssef Tarek Ichim, Oana Shetty, Pranav Wang, Dongsheng Ma, Zhiqiang Nourbakhsh, Armineh Liu, Xiaomo |
| contents | Due to their ability to process long and complex contexts, LLMs can offer key benefits to the Legal domain, but their adoption has been hindered by their tendency to generate unfaithful, ungrounded, or hallucinatory outputs. While Retrieval-Augmented Generation offers a promising solution by grounding generations in external knowledge, it offers no guarantee that the provided context will be effectively integrated. To address this, context-aware decoding strategies have been proposed to amplify the influence of relevant context, but they usually do not explicitly enforce faithfulness to the context. In this work, we introduce Confidence-guided Copy-based Decoding for Legal Text Generation (CoCoLex)-a decoding strategy that dynamically interpolates the model produced vocabulary distribution with a distribution derived based on copying from the context. CoCoLex encourages direct copying based on the model's confidence, ensuring greater fidelity to the source. Experimental results on five legal benchmarks demonstrate that CoCoLex outperforms existing context-aware decoding methods, particularly in long-form generation tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_05534 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | CoCoLex: Confidence-guided Copy-based Decoding for Grounded Legal Text Generation S, Santosh T. Y. S. Elkhayat, Youssef Tarek Ichim, Oana Shetty, Pranav Wang, Dongsheng Ma, Zhiqiang Nourbakhsh, Armineh Liu, Xiaomo Computation and Language Due to their ability to process long and complex contexts, LLMs can offer key benefits to the Legal domain, but their adoption has been hindered by their tendency to generate unfaithful, ungrounded, or hallucinatory outputs. While Retrieval-Augmented Generation offers a promising solution by grounding generations in external knowledge, it offers no guarantee that the provided context will be effectively integrated. To address this, context-aware decoding strategies have been proposed to amplify the influence of relevant context, but they usually do not explicitly enforce faithfulness to the context. In this work, we introduce Confidence-guided Copy-based Decoding for Legal Text Generation (CoCoLex)-a decoding strategy that dynamically interpolates the model produced vocabulary distribution with a distribution derived based on copying from the context. CoCoLex encourages direct copying based on the model's confidence, ensuring greater fidelity to the source. Experimental results on five legal benchmarks demonstrate that CoCoLex outperforms existing context-aware decoding methods, particularly in long-form generation tasks. |
| title | CoCoLex: Confidence-guided Copy-based Decoding for Grounded Legal Text Generation |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2508.05534 |