Prefix-Tuning for Arbitrary Output Sequences on Pretrained Transformers
arXiv:2610.24123v1 Announce Type: new Abstract: Prefix-tuning is a parameter-efficient fine-tuning method for large language models (LLMs) that steers model outputs by prepending task-specific prefix representations to the input sequence. Despite its practical success, the theoretical understanding…
Read the full story at arXiv cs.CL ↗
Timeline · 1 report
- 2026-10-20 04:00 · arXiv cs.CL
Prefix-Tuning for Arbitrary Output Sequences on Pretrained Transformers