Length-regularized self-distillation trains rerankers to use concise reasoning traces that match teacher performance at 34-37 percent lower token cost.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Stop Overthinking: Unlocking Efficient Listwise Reranking with Minimal Reasoning
Length-regularized self-distillation trains rerankers to use concise reasoning traces that match teacher performance at 34-37 percent lower token cost.