A Memorization Floor for LLM Refinement of Decompiled Code

• URL: http://arxiv.org/abs/2609.17236v1
• Date: Tue, 15 Sep 2026 14:15:08 GMT
• Title: A Memorization Floor for LLM Refinement of Decompiled Code
• Abstract summary: We introduce a within-item control separating what LLM output recovers from its input from what it recovers from its prior.<n>Because the comparison is within-item, corpus difficulty cannot contribute; it costs twenty API calls.<n> refiner from another vendor reproduces this -- twelve contrasts, two models, twelve nulls.<n>No registered hypothesis was confirmed, and we report the five instrument failures behind that in full.
• Score: 0.0
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We introduce a memorization floor: a within-item control separating what LLM refinement of decompiler output recovers from its input from what it recovers from its prior. Refine a function, then refine it again from an input whose identifiers have been destroyed, and measure what survives. Because the comparison is within-item, corpus difficulty cannot contribute; it costs twenty API calls. Applied to functions written after our analysis plan was committed, so no released model could have memorized them, it reports two things. Recovery is real: refined output sits +0.072 to +0.137 above an arm-matched permutation null built from its own output vocabulary. But it does not depend on the input we ablate: destroying the input's dataflow changes the naming gain by +0.001 (95% CI [-0.026, +0.026]), and removing type prefixes or permuting names changes it by no more. A second refiner from another vendor, registered in advance and given byte-identical inputs, reproduces this -- twelve contrasts, two models, twelve nulls. Readability stays at ceiling throughout, so a reader is given no signal. The null is bounded, not absolute: contributions under 0.056 are invisible, and the ablation leaves operations intact, so naming from those alone remains a competing reading. No registered hypothesis was confirmed, and we report the five instrument failures behind that in full, including a reassembly harness biased against the treated arm and an equivalence checker we registered without checking it worked on our inputs.

Related papers

• Clean Engineering, Unstable Measurement: A Preregistered Reliability Failure of Black-Box LLM Observers on Shared Endpoints [4.41537705949485]
  Language-model judges now gate training data, score generations, and drive leaderboards.<n>We audited that assumption in two preregistered campaigns with every threshold fixed in advance.<n>Across 52,988 audited request attempts, same-window repeat rankings agreed at Spearman 0.400 against a required 0.90.<n>We distill the evidence into a three-level snapshot-identity ladder, eight design rules, and a reporting checklist.
  arXiv  Detail & Related papers  (2026-09-03T17:59:43Z)
• VeriPhy: Agentic Physical Reasoning for World Model Evaluation and Refinement [57.86962208273888]
  We present an auditable physical-verification system in which a text-only planner compiles the prompt into typed physical obligations.<n>Each action returns a provenance-carrying evidence record whose payload, when usable, is either a typed measurement or an explicitly tagged learned state.<n>We anchor evaluation in a 1,500-clip corpus of human-annotated flaw records that localize real generation failures in prompt reference, space, and time.
  arXiv  Detail & Related papers  (2026-09-02T20:36:45Z)
• Feedback That Backfires: Why Small Language Model Agents Repeat the Call They Just Watched Fail [0.0]
  Defining the corrective gain of a failure record as the change in log-probability of re-emitting the action that just failed, we find the gain is negative for every instruction-tuned model.<n>The problem is in the harness, not the model's grasp of error messages, and that predicts which remedies work.
  arXiv  Detail & Related papers  (2026-08-24T10:19:17Z)
• The Recall Trap: A Recall-Maximizing Retriever Configuration Reduces Issue Resolution in Fixed-Budget Code Context [45.88028371034407]
  We report a controlled case study in code repair, not a new phenomenon.<n>We toggle one flag (one-chunk-perfile deduplication) on an otherwise identical stack.<n>We map where it holds: it reverses on a lexical BM25 (3.2pp, significant cross-paradigm interaction)<n>We show that at a tight fixed budget, do not hard-deduplicate by file, and A/B packing policies against the task, not the metric flag was tuned.
  arXiv  Detail & Related papers  (2026-08-14T19:22:50Z)
• Repair, Not Improvement: Decomposing Constrained Decoding in Tool-Call Abstention [0.0]
  We show how an enum leaves the wording of an answer alone and narrows the set of answers there are, and declining to call anything is the first it drops.<n>We evaluate open-weight models from 0.6B to 4B on matched English and Korean items, so the language comparison is made within item.<n>What it recovers is form: of 698 abstentions repaired, 545 had no readable answer and 0 were the scorer refused.
  arXiv  Detail & Related papers  (2026-08-14T04:58:01Z)
• Similarity Gates Approve Reversals: A Validity Audit of Embedding-Cosine Thresholds in Agent Systems [0.0]
  Agent frameworks ship quality gates that compare text blocks by embedding-cosine similarity and decide at a fixed cutoff.<n>Deduplication filters, semantic caches, drift guards, and answer grader gates deploy to answer the question: "Does this text still mean the same thing?"<n>We audit this gate class as a measurement instrument.
  arXiv  Detail & Related papers  (2026-08-10T20:36:40Z)
• The Announcement Carries the Cue: Markup, Boundaries, and the Notation of Pre-Training Corpora [0.0]
  How a document's arrangement is written down, its notation, is a training variable that no dataset card records.<n>We define clean-window survival, a deterministic count of how much of a stream still demands the boundary inference, and measure notation on three fronts.
  arXiv  Detail & Related papers  (2026-08-10T03:43:30Z)
• Which Defense Closes Which Threat? Attributing OWASP-LLM-Top-10 Coverage and Its Brittleness Under Paraphrasing [51.56484100374058]
  Production LLM applications stack several defense families -- refusal-phrase filters, tokenbudget controls, model allowlists, rate limits, tool-registry authentication -- yetBAS benchmarks report a single aggregate coverage number.<n>We add four-LLM-Top-10-aware agents to a 21-agent baseline scanner and target a lattice of four synthetic LLM endpoints.
  arXiv  Detail & Related papers  (2026-06-01T19:39:25Z)
• Skill-as-Pseudocode: Refactoring Skill Libraries to Pseudocode for LLM Agents [50.78646963535282]
  We propose Skill-as-Pseudocode (SaP), an automatic conversion of markdown skill libraries into typed pseudocode.<n>For each cluster of similar procedural passages drawn from one or more skills, SaP extracts a typed contract and filters it through a four-check deterministic verifier.<n>Promoted contracts are inlined into a rewritten skill skeleton together with restored concrete action templates, giving the agent two complementary signals.
  arXiv  Detail & Related papers  (2026-05-27T04:48:40Z)
• From Early Encoding to Late Suppression: Interpreting LLMs on Character Counting Tasks [49.57538588967748]
  Large language models (LLMs) exhibit failures on elementary symbolic tasks such as character counting in a word, despite excelling on complex benchmarks.<n>We uncover a consistent phenomenon across modern architectures, including LLaMA, Qwen, and Gemma.<n>We show that symbolic reasoning failures in LLMs are not due to missing representations or insufficient scale, but arise from structured interference within the model's computation graph.
  arXiv  Detail & Related papers  (2026-04-01T11:40:12Z)
• D-LiFT: Improving LLM-based Decompiler Backend via Code Quality-driven Fine-tuning [49.16469288280772]
  Decompilers reconstruct human-readable source code from binaries.<n>Despite recent advances, their outputs often suffer from syntactic and semantic errors and remain difficult to read.<n>With the advent of large language models (LLMs), researchers began to explore the potential of LLMs to refine decompiler output.<n>We present D-LIFT, an enhanced decompiler-LLM pipeline with fine-tuned reinforcement learning.
  arXiv  Detail & Related papers  (2025-06-11T19:09:08Z)
• Can LLMs Separate Instructions From Data? And What Do We Even Mean By That? [60.50127555651554]
  Large Language Models (LLMs) show impressive results in numerous practical applications, but they lack essential safety features.<n>This makes them vulnerable to manipulations such as indirect prompt injections and generally unsuitable for safety-critical tasks.<n>We introduce a formal measure for instruction-data separation and an empirical variant that is calculable from a model's outputs.
  arXiv  Detail & Related papers  (2024-03-11T15:48:56Z)
• Composed Fine-Tuning: Freezing Pre-Trained Denoising Autoencoders for Improved Generalization [93.95299500688286]
  We focus on prediction problems with structured outputs subject to output validity constraints. We propose composed fine-tuning, which fine-tunes a predictor composed with the pre-trained denoiser. For two-layer ReLU networks, we prove that composed fine-tuning significantly reduces the complexity of the predictor.
  arXiv  Detail & Related papers  (2020-06-29T17:14:35Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.