Information-Theoretic and Prompt-Based Evaluation of Discourse Connective Edits in Instructional Text Revisions

Berfin Aktas, Michael Roth


Abstract
We present a dataset of text revisions involving the deletion or replacement of discourse connectives. Manual annotation of a replacement subset reveals that only 19% of edits were judged either necessary or should be left unchanged, with the rest appearing optional. Surprisal metrics from GPT-2 token probabilities and prompt-based predictions from GPT-4.1 correlate with these judgments, particularly in such clear cases.
Anthology ID:
2025.codi-1.19
Volume:
Proceedings of the 6th Workshop on Computational Approaches to Discourse, Context and Document-Level Inferences (CODI 2025)
Month:
November
Year:
2025
Address:
Suzhou, China
Editors:
Michael Strube, Chloe Braud, Christian Hardmeier, Junyi Jessy Li, Sharid Loaiciga, Amir Zeldes, Chuyuan Li
Venues:
CODI | WS
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
228–236
Language:
URL:
https://aclanthology.org/2025.codi-1.19/
DOI:
Bibkey:
Cite (ACL):
Berfin Aktas and Michael Roth. 2025. Information-Theoretic and Prompt-Based Evaluation of Discourse Connective Edits in Instructional Text Revisions. In Proceedings of the 6th Workshop on Computational Approaches to Discourse, Context and Document-Level Inferences (CODI 2025), pages 228–236, Suzhou, China. Association for Computational Linguistics.
Cite (Informal):
Information-Theoretic and Prompt-Based Evaluation of Discourse Connective Edits in Instructional Text Revisions (Aktas & Roth, CODI 2025)
Copy Citation:
PDF:
https://aclanthology.org/2025.codi-1.19.pdf
Supplementarymaterial:
 2025.codi-1.19.SupplementaryMaterial.zip