A contrastive learning framework with positive and negative example pairs improves faithfulness and slightly reduces toxicity on Reddit TL;DR summarization, but the central claim of avoiding the alignment tax is not evaluated on the paper's own knowledge benchmarks.
Title resolution pending
1 Pith paper cite this work, alongside 131 external citations. Polarity classification is still indexing.
1
Pith paper citing it
131
external citations · OpenAlex
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Paying Alignment Tax with Contrastive Learning
A contrastive learning framework with positive and negative example pairs improves faithfulness and slightly reduces toxicity on Reddit TL;DR summarization, but the central claim of avoiding the alignment tax is not evaluated on the paper's own knowledge benchmarks.