Pith. sign in

REVIEW 8 cited by

Security Vulnerability Detection with Multitask Self-Instructed Fine-Tuning of Large Language Models

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2406.05892 v1 pith:NGVGQZ7E submitted 2024-06-09 cs.CR cs.LGcs.SE

Aidan Z.H. Yang , Haoye Tian , He Ye , Ruben Martins , Claire Le Goues This is my paper · ORCID
classification cs.CRcs.LGcs.SE
keywords vulnerabilitydetectionmodelscodedatalanguagemsivdmultitask
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Software security vulnerabilities allow attackers to perform malicious activities to disrupt software operations. Recent Transformer-based language models have significantly advanced vulnerability detection, surpassing the capabilities of static analysis based deep learning models. However, language models trained solely on code tokens do not capture either the explanation of vulnerability type or the data flow structure information of code, both of which are crucial for vulnerability detection. We propose a novel technique that integrates a multitask sequence-to-sequence LLM with pro-gram control flow graphs encoded as a graph neural network to achieve sequence-to-classification vulnerability detection. We introduce MSIVD, multitask self-instructed fine-tuning for vulnerability detection, inspired by chain-of-thought prompting and LLM self-instruction. Our experiments demonstrate that MSIVD achieves superior performance, outperforming the highest LLM-based vulnerability detector baseline (LineVul), with a F1 score of 0.92 on the BigVul dataset, and 0.48 on the PreciseBugs dataset. By training LLMs and GNNs simultaneously using a combination of code and explanatory metrics of a vulnerable program, MSIVD represents a promising direction for advancing LLM-based vulnerability detection that generalizes to unseen data. Based on our findings, we further discuss the necessity for new labelled security vulnerability datasets, as recent LLMs have seen or memorized prior datasets' held-out evaluation data.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 8 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Geometric quantification for nonlinear deformation in knitted fabrics

    cond-mat.soft 2026-04 unverdicted novelty 7.0 of 10

    A geometric quantification framework reconstructs yarn centerlines and fabric surfaces from sparse knit data and partitions large deformation into stitch reorientation, loop bending, surface bending, and dilation.

  2. DREA: Decoupled Reasoning and Exploration Agents for Repository-Level Vulnerability Detection

    cs.CR 2026-07 conditional novelty 6.0 of 10

    DREA improves repository-level vulnerability detection by coupling an LLM planner that forms hypotheses with a cheap local explorer that gathers cross-file evidence, lifting paired accuracy from 19-26% to 30-42% at mu...

  3. VulWeaver: Weaving Broken Semantics for Grounded Vulnerability Detection

    cs.SE 2026-04 unverdicted novelty 6.0 of 10

    VulWeaver improves Java vulnerability detection to 0.75 F1 by enhancing dependency graphs with LLM semantic fixes, extracting full context from slices plus implicit usage info, and applying type-specific meta-promptin...

  4. VULPO: Context-Aware Vulnerability Detection via On-Policy LLM Optimization

    cs.CR 2025-11 conditional novelty 6.0 of 10

    A 4B LLM trained with judge-scored, difficulty-scaled on-policy RL reaches F1 70.45 on context-aware C/C++ vulnerability detection, comparable to DeepSeek-V3.1, with the caveat that the same judge provided the trainin...

  5. LLMxCPG: Context-Aware Vulnerability Detection Through Code Property Graph-Guided Large Language Models

    cs.CR 2025-07 conditional novelty 6.0 of 10

    LLMxCPG uses LLM-generated CPG queries to slice code down to vulnerability-relevant paths and a fine-tuned classifier to label those slices, outperforming several baselines on function-level and project-level datasets.

  6. A Mixture of Linear Corrections Generates Secure Code

    cs.CR 2025-07 conditional novelty 6.0 of 10

    An inference-time mixture of linear correction vectors, derived from linear probes on LLM hidden states, improves the security and functionality of code generated by Qwen2.5-Coder and CodeLlama models.

  7. Improving LLM Reasoning for Vulnerability Detection via Group Relative Policy Optimization

    cs.CR 2025-07 conditional novelty 6.0 of 10

    GRPO fine-tuning with a dynamic formatting, reasoning, and correctness reward helps small LLMs detect vulnerabilities more accurately and reason more coherently than supervised fine-tuning.

  8. Evaluating Language Models For Threat Detection in IoT Security Logs

    cs.CR 2025-07 conditional novelty 4.0 of 10

    Fine-tuned open-source LLMs outperform classical ML classifiers on multiclass attack classification in the Edge-IIoTset dataset, while the mitigation-quality claims rest on a self-referential evaluation.

Pith tools