SandRise
Ep 959 Research Paper 0:58 w/ Onyx

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

We study over-editing, the tendency of a model to rewrite code beyond what is required to fix a bug. We construct an evaluation framework from 400 BigCodeBench problems by injecting controlled AST-level corruptions into reference solutions, giving each repair task a known minimal patch.

Research Paper
Read the source →
Embed this episode

Paste this on any site — the player is a self-contained iframe with no cookies or trackers. 

<iframe src="https://sandrise.io/exploring-next/embed/959"
  width="100%" height="180" style="max-width:640px;border:0;border-radius:12px;overflow:hidden"
  title="Exploring Next — Episode 959 audio player"
  loading="lazy" allow="autoplay" referrerpolicy="strict-origin-when-cross-origin"></iframe>
Embed & API docs →
Script Built In Brief Voice ElevenLabs v3

Transcript

Onyx When Models Edit Too Much: On the Fidelity of Minimal Code Edits. Worth a couple minutes.

Onyx we study over-editing, the tendency of a model to rewrite code beyond what is required to fix a bug.

Onyx What is interesting here is our results show that edit fidelity is a distinct axis of code-repair quality: (1) Over-editing is widespread: several frontier models, such as GPT-5.5, achieve competitive Pass@1 while still making unnecessarily large edits.

Onyx reasoning models also raise a code-specific constraint-following question: although reasoning often improves coding and instruction-following performance ( Zhou et al., 2023 ) , recent work shows that stronger reasoning can interact poorly with explicit constraints ( Fu et al..

Onyx If you want to go deeper, the source is worth reading — link's in the show notes.

Onyx That is the brief.