arXiv Machine Learning By Andikawati P Widjaja, Yongjun Kim, Hyounghun Kim, Jaeho Lee

PARTREP: Learning What to Repeat for Decoder-only LLMs

Read the original on arXiv Machine Learning →

arXiv:2607. 01792v1 Announce Type: cross Abstract: While decoder-only LLMs excel at a vast array of natural language tasks, it suffers from an asymmetric information flow induced by causal attention: later tokens are richer in contextual grounding than earlier ones.

Summary generated by The Flow from the publisher's feed. The full article lives at arXiv Machine Learning.

Hugging Face Trending Papers
Jul 2

PARTREP: Learning What to Repeat for Decoder-only LLMs

While decoder-only LLMs excel at a vast array of natural language tasks, it suffers from an asymmetric information flow induced by causal attention: later tokens are richer in contextual grounding than earlier ones. A simple and effective remedy is prompt repetition -- just appending a second copy of prompt before generation can redistribute grounding across positions and improve reasoning performance.