Hugging Face Trending Papers

PARTREP: Learning What to Repeat for Decoder-only LLMs

Read the original on Hugging Face Trending Papers →

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.

Summary generated by The Flow from the publisher's feed. The full article lives at Hugging Face Trending Papers.