Token Constraint Decoding Improves Robustness on Question Answering for Large Language Models
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866916790540435456 |
|---|---|
| author | Yao, Jui-Ming Chen, Hao-Yuan Tang, Zi-Xian Tan, Bing-Jia Peng, Sheng-Wei Xie, Bing-Cheng Su, Shun-Feng |
| author_facet | Yao, Jui-Ming Chen, Hao-Yuan Tang, Zi-Xian Tan, Bing-Jia Peng, Sheng-Wei Xie, Bing-Cheng Su, Shun-Feng |
| contents | Large Language Models (LLMs) have demonstrated impressive performance on multiple-choice question answering (MCQA) benchmarks, yet they remain highly vulnerable to minor input perturbations. In this paper, we introduce and evaluate Token Constraint Decoding (TCD). This simple yet effective inference-time algorithm enforces alignment between token-level predictions to enhance robustness in noisy settings. Through extensive experiments on CommonsenseQA, MMLU, and MMLU-Pro, we show that TCD, especially when paired with prompt engineering (PE) fixes, significantly restores performance degraded by input noise, yielding up to +39\% absolute gains for weaker models like Gemma3 1B. Penalty sweep analyses further reveal that TCD implicitly regularizes overconfident outputs, with different models requiring distinct penalty schedules to maximize resilience. Our findings establish TCD as a practical, model-agnostic approach for improving reasoning stability under real-world imperfections and pave the way for more reliable deployment of LLMs in safety-critical or user-facing applications. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_09408 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Token Constraint Decoding Improves Robustness on Question Answering for Large Language Models Yao, Jui-Ming Chen, Hao-Yuan Tang, Zi-Xian Tan, Bing-Jia Peng, Sheng-Wei Xie, Bing-Cheng Su, Shun-Feng Computation and Language Artificial Intelligence Large Language Models (LLMs) have demonstrated impressive performance on multiple-choice question answering (MCQA) benchmarks, yet they remain highly vulnerable to minor input perturbations. In this paper, we introduce and evaluate Token Constraint Decoding (TCD). This simple yet effective inference-time algorithm enforces alignment between token-level predictions to enhance robustness in noisy settings. Through extensive experiments on CommonsenseQA, MMLU, and MMLU-Pro, we show that TCD, especially when paired with prompt engineering (PE) fixes, significantly restores performance degraded by input noise, yielding up to +39\% absolute gains for weaker models like Gemma3 1B. Penalty sweep analyses further reveal that TCD implicitly regularizes overconfident outputs, with different models requiring distinct penalty schedules to maximize resilience. Our findings establish TCD as a practical, model-agnostic approach for improving reasoning stability under real-world imperfections and pave the way for more reliable deployment of LLMs in safety-critical or user-facing applications. |
| title | Token Constraint Decoding Improves Robustness on Question Answering for Large Language Models |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2506.09408 |