LLMs learn length biases from the examples in their prompt, and rebalancing those examples can offset a length bias created by finetuning.
Title resolution pending
1 Pith paper cite this work, alongside 67 external citations. Polarity classification is still indexing.
1
Pith paper citing it
67
external citations · OpenAlex
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
In-Context Learning (and Unlearning) of Length Biases
LLMs learn length biases from the examples in their prompt, and rebalancing those examples can offset a length bias created by finetuning.