← Back to all news
arXiv AI By Priyansh Srivastava, Romit Chatterjee

Suppressed, Not Erased: A Representational Trace of Edited Facts Survives Even Weight-Free Knowledge Editing

Read the original on arXiv AI →

The paper investigates whether knowledge editing truly erases original facts from language models. Using a linear trace probe, the authors find that after editing a fact in GPT‑2‑XL, the original object remains highly decodable from hidden states across three different editing methods, even when the model behaves correctly on edited prompts. This suggests that editing suppresses rather than removes the original association in representational space.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

Related stories

arXiv AI
Sep 25

Sequential knowledge editing breaks a model's ability to tell good evidence from bad, without costing it accuracy

The paper investigates how sequential knowledge editing can degrade a language model’s ability to discern reliable evidence from unreliable evidence without affecting overall accuracy. Using a conservatively tuned LoRA on Qwen2.5‑7B‑Instruct, the authors show that after 1,000 edits the model’s arbitration score for untouched facts drops by 36%, leading to higher error rates on its most confident decisions, while MMLU accuracy remains unchanged. The study also finds that in some model‑method combinations, sequential edits can reduce MMLU to chance levels even though edit success and locality remain perfect.

By Atul Anand
fine-tuningbenchmarks
More like this →
arXiv AI
Sep 4

When Models Edit Too Much: On the Fidelity of Minimal Code Edits

The paper investigates the problem of over‑editing by large language models when repairing code, showing that even state‑of‑the‑art models like GPT‑5.5 frequently rewrite more code than necessary. Using a benchmark of 400 BigCodeBench problems with controlled AST corruptions, the authors quantify excess edits and demonstrate that a simple preservation instruction can reduce unnecessary changes and improve pass rates. They further explore training strategies, finding that reinforcement learning yields the best balance between edit fidelity and performance retention, highlighting edit fidelity as a distinct, measurable dimension of code‑repair quality.

By Tongyao Zhu, Wei Hern Lim, Min-Yen Kan
llmsreinforcement-learningfine-tuning
More like this →
arXiv AI
Sep 17

Contiguity, Not Importance: Budgeted Repair of Stale KV Caches After Document Edits

The paper investigates how to efficiently repair stale key-value (KV) caches in retrieval‑augmented generation systems after document edits. It proposes a budgeted in‑place recomputation approach and evaluates training‑free position‑selection policies on a factual RAG benchmark. Across three model families, a contiguous edit‑local window consistently recovers most of the post‑edit answer quality while being 13–21 times faster than a full re‑prefill, though its effectiveness diminishes when answer‑bearing text moves downstream.

By Mingyang Mao, Wyatt Mackey, Xiaomin Lin
ragagentsbenchmarks
More like this →