@inproceedings{yadav-2026-competence,
    title = "Competence Collapse in Code-Mixed Generation: Spectral Evidence and Mechanistic Recovery via Cross-Lingual Activation Steering",
    author = "Yadav, Tanushree Ravindra Pratap",
    editor = "Hettiarachchi, Hansi  and
      Ranasinghe, Tharindu  and
      Plum, Alistair  and
      Rayson, Paul  and
      Mitkov, Ruslan  and
      Gaber, Mohamed  and
      Premasiri, Damith  and
      Tan, Fiona Anting  and
      Uyangodage, Lasitha",
    booktitle = "Proceedings of the Second Workshop on Language Models for Low-Resource Languages ({L}o{R}es{LM} 2026)",
    month = mar,
    year = "2026",
    address = "Rabat, Morocco",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/manual-author-scripts/2026.loreslm-1.3/",
    pages = "29--40",
    ISBN = "979-8-89176-377-7",
    abstract = "As Large Language Models (LLMs) approachhuman-level reasoning in English, their performance in low-resource, code-mixed languagesremains surprisingly brittle. We identify Competence Collapse, a distinct pathology wheremodels capable of complex reasoning in English exhibit severe utility degradation whenprompted in Hinglish (Hindi-English). Wequantify this as a Service Gap, observing astatistically significant decline in instructionalquality ({\ensuremath{\Delta}}D {\ensuremath{\approx}} {\ensuremath{-}}11.3{\%}, p {\ensuremath{<}} 0.001) across9 diverse architectures. Spectral analysis suggests that this stems from a representationaldivergence between the model{'}s High-UtilityDirection and its Generation Subspace. Tobridge this gap, we propose Cross-LingualActivation Steering (CLAS), an inferencetime intervention that injects a ``CompetenceGap Vector'' into the residual stream. Evaluated across 6 open-weight models (using alightweight calibration set, N = 50), CLASrecovered utility by {\ensuremath{\Delta}}D = +2.22 (d = 0.60)while preserving code-mixed fidelity (CMI {\ensuremath{\approx}}0.4) and reinforcing safety protocols."
}