Weighting the EOS token in the loss during fine-tuning reduces too-long summaries on CNN/DailyMail and fixed-length XL-sum, but not on dynamic-length XL-sum.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Controlling Summarization Length Through EOS Token Weighting
Weighting the EOS token in the loss during fine-tuning reduces too-long summaries on CNN/DailyMail and fixed-length XL-sum, but not on dynamic-length XL-sum.