Multi-amateur contrastive decoding pools signals from a set of small language models, with mean or consensus aggregation, to improve open-ended text generation over single-amateur CD.
Exploring the limits of transfer learning with a unified text-to-text transformer,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Multi-Amateur Contrastive Decoding for Text Generation
Multi-amateur contrastive decoding pools signals from a set of small language models, with mean or consensus aggregation, to improve open-ended text generation over single-amateur CD.