Reproduce “Decoding by Contrasting Knowledge: Enhancing LLMs' Confidence on Edited Facts” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Baolong Bi, Shenghua Liu, Lingrui Mei, Yiwei Wang, Pengliang Ji, Xueqi Cheng
DeCK uses contrasting predictions from an edited context and the original question to guide next-token decoding without changing model weights. It addresses “stubborn knowledge,” where an edit raises support for a new fact but the old answer remains more likely. On LLaMA3-8B-Instruct and MQuAKE-HARD, accuracy rises from 14.3% to 45.7% in the reported batch-one evaluation, not establishing universal gains.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.