llm-based-repository-level-program-repair-telling-generator-be-minimal-or CLM-0589mechanismsingle paperReviewed by AI
In LLM-based repository-level program repair, telling the generator to be minimal or constraining its output format does not shrink patches, because verbosity comes from the capability-raising machinery itself (iterative refinement, broad context); a separate post-generation refiner that rewrites an already-plausible patch is what reduces size without losing fixes.
Capability: Cutting to the chase
Observed on
Repository-level issue fixing measured on SWE-bench Verified; patch size measured relative to the developer patch across 28 leaderboard systems and four host agents..
Sources
- Validated by: On SWE-bench Verified across four host systems (Agentless, SWE-agent, Moatless, OpenHands), average total changes relative to developer patches fell from +242.14% to +4.24% and net changes from +348.24% to -39.75%, while resolved instances were preserved or improved by up to +42; prompting Claude-4.5-Sonnet, Atomizer untangling, AdaPatcher and PRepair all shrank patches only by losing 29-217 resolved instances. Ablation: SFT alone over-compresses (resolves 276 avg, below the untrained 293); DPO restores 332.. Measured. [truncated]
Either way, bring evidence: a link, or what you saw.
Disagreeing is the most useful thing you can do here. Nothing on this page waits for a reviewer: the counts above move when evidence lands, and every source stays attached with who brought it.
Related claims
- On repository-level issue-fixing tasks (SWE-bench Lite/Verified/Pro), spending effort upstream — expanding the issue text into a structured requirement spec from retrieved repo context, then testing and refining that spec via generated code and tests — raises resolve rate over agents that consume the raw issue description, and raises the share of patches that even apply cleanly by a larger margin than it raises correctness.Generating and editing working code · reviewed by AI
- For LLM program-repair workflows built on self-correction, generated-knowledge or multi-agent scaffolding, the gain over plain prompting is unstable across benchmarks and can turn negative once extrinsic defect information (error messages, triggering tests, bug comments) is withheld, while simply putting the error message plus the failing test into the prompt yields more correct repairs than the scaffolds do.Fixing its own mistakes · reviewed by AI
- On repository-level issue-fixing tasks (SWE-bench Lite/Verified/Pro), spending effort upstream — expanding the issue text into a structured requirement spec from retrieved repo context, then testing and refining that spec via generated code and tests — raises resolve rate over agents that consume the raw issue description, and raises the share of patches that even apply cleanly by a larger margin than it raises correctness.Generating and editing working code · reviewed by AI
- In LLM-based program repair on Defects4J bugs with frontier models (GPT-5, DeepSeek-R1, Claude Sonnet 4.5) prompted zero-shot, passing the developer-written test suite does not mean the patch is grounded in the bug: manual inspection of 812 sampled repairs found ungrounded repairs in 72.7% of cases including test-passing patches, most often from incorrect causal localization (45.9%) rather than a wrong edit style, and the same models frequently misidentify the bug-triggering tests and generate additional tests that miss the triggering condition.Stating false facts confidently · reviewed by AI
- In search-based automated program repair of Simulink/Stateflow models, replacing hand-crafted domain-specific mutation operators with an LLM that proposes label rewrites from static component context lowers repair coverage under a fixed wall-clock budget, because unconstrained generative edits widen the candidate space while each candidate costs an expensive simulation and the generator gets no behavioural feedback about why a patch failed.Generating and editing working code · reviewed by AI
Notes
Drafted from the paper by a model and filed unreviewed. Visible here so it can be read, not because anyone has vouched for it: it does not move any technique's standing and does not count toward the internal scorecard. Drafted confidence: medium. Falsifier as drafted: A repair system whose patches match developer patch size after only a minimality instruction or output-format change, with resolution unchanged; or a post-generation refiner that shrinks patches no more than a minimality prompt at matched resolution.