CORTEXA
← Browse
arxivcs.SEcs.LG2026-07-14

Toward Localizing and Repairing Bias in Transformer Attention Heads

Sigma Jahan

Transformer language models are increasingly used as software components, yet biased outputs remain difficult to localize and repair inside the model. Existing fairness testing and repair methods largely operate at the input-output or retraining level, while recent work suggests that bias-related behavior can concentrate in a small set of attention heads. This paper studies whether attention heads can be localized and repaired through a targeted inference-time intervention. We introduce ROBIN, a white-box head-level fairness debugging method that ranks attention heads using sensitivity to fairness probes and removes a small bias subspace from selected head outputs. In a four-model pilot study, ROBIN reduces the measured WinoBias gap across all models while preserving language-modeling quality better than whole-head zeroing. These preliminary results suggest that head-level bias repair should consider not only which heads are selected, but also how selected heads are modified.

View free PDFSource page

Related papers

arxivcs.SEcs.LG2026-07-13

TraceSynth: Generating Production-Quality Kernel Traces with Constraint-Guided Diffusion Models

Yuvraj Sehgal, Sneh Patel, Mahsa Panahandeh, Naser Ezzati-Jivan, Francois Tetreault

Machine learning models for system diagnostics rely on kernel execution traces to capture fine-grained system behavior, but collecting production traces in industrial systems is costly due to runtime overhead, storage demands, and privacy constraints. We present TraceSynth, a dif…

View free PDFSource page
arxivcs.SEcs.AIcs.LG2026-07-14

Form, Not Content? A Preregistered, Placebo-Controlled Evaluation of Learned Error-Conditioned Self-Repair Through Prompts and Weights in Frozen Small Code Models

Mehmet Iscan

Frozen small code LLMs are deployed locally, yet the information guiding a retry after a failed attempt is still measured without placebo controls in the self-repair literature. We treat a failed program as a conjecture and an execution counterexample as an oracle-relative refuta…

View free PDFSource page
arxivcs.CLcs.LGcs.SE2026-07-18

OpenLanguageModel: Readable and Composable Small-Language-Model Pretraining for Education and Research

Tavish Mankash, Vardhaman Kalloli, Keshava Prasad, Deepan Muthirayan

OpenLanguageModel (OLM) is an open-source PyTorch library for building and pretraining small language models while keeping their machinery visible. In OLM, model code reads like the architecture: components are ordinary modules, while Block, Residual, Repeat, and Parallel describ…

View free PDFSource page
arxivcs.SEcs.LG2026-07-11

Toward Production-Ready Federated Learning in Healthcare: Privacy, Orchestration, and Governance in MLOps

Sakshi Gorkhali, Jonesh Shrestha

Healthcare organizations often cannot freely centralize patient data because medical records are sensitive, regulated, and institutionally controlled. Federated learning offers a practical alternative by allowing hospitals and clinics to train a shared model while keeping raw dat…

View free PDFSource page
arxivcs.LGcs.SE2026-07-11

Knowledge-Conditioned, Single-Pass LLM Synthesis of Executable Unity Game Scenes: A Compiler Error Census across 26 Goal Playable Concepts

Hugh Xuechen Liu, Kıvanç Tatar

Large language models (LLMs) write Unity C\# for game scenes. Yet nearly all demonstrations rest on an iterative repair loop that regenerates code until it compiles, conflating what the model writes with what the loop fixes. We remove the loop and evaluate a single pass, where th…

View free PDFSource page
arxivcs.SEcs.CRcs.LG2026-06-30

An Empirical Study of Security Calibration in Large Language Models for Code

Mohammed Latif Siddiq, Md. Nafiu Rahman, Joanna C. S. Santos

Large Language Models (LLMs) are rapidly transforming software development, yet their use in security-critical contexts raises a key question: do models know when their generated code is insecure? This property, known as calibration, measures whether a model's confidence aligns w…

View free PDFSource page