RoParQ: Paraphrase-Aware Alignment of Large Language Models Towards Robustness to Paraphrased Questions
Fuente:
arXiv
Gespeichert in:
| 1. Verfasser: | |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915639485005824 |
|---|---|
| author | Choi, Minjoon |
| author_facet | Choi, Minjoon |
| contents | Large Language Models (LLMs) often exhibit inconsistent behavior when answering paraphrased questions, suggesting a reliance on surface-level patterns rather than true semantic understanding. To address this limitation, we introduce RoParQ, a benchmark specifically constructed to evaluate cross-paraphrase consistency in closed-book multiple-choice QA. This benchmark is derived from standard datasets by generating paraphrases via proprietary models and selectively retaining examples that elicit inconsistent confidence from a judge model. We further propose XParaCon, a novel evaluation metric that quantifies a model's robustness by measuring the standard deviation of accuracies across question variants. Additionally, we implement a reasoning-based, paraphrase-aware Supervised Fine-Tuning (SFT) strategy designed to align models toward semantic invariance. Our experiments demonstrate that this targeted alignment significantly enhances robustness. Notably, fine-tuned lightweight models achieved consistency levels comparable to much larger pre-trained models. These results highlight the efficacy of our approach in mitigating superficial memorization and fostering more robust, reliable LLMs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_21568 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | RoParQ: Paraphrase-Aware Alignment of Large Language Models Towards Robustness to Paraphrased Questions Choi, Minjoon Computation and Language Large Language Models (LLMs) often exhibit inconsistent behavior when answering paraphrased questions, suggesting a reliance on surface-level patterns rather than true semantic understanding. To address this limitation, we introduce RoParQ, a benchmark specifically constructed to evaluate cross-paraphrase consistency in closed-book multiple-choice QA. This benchmark is derived from standard datasets by generating paraphrases via proprietary models and selectively retaining examples that elicit inconsistent confidence from a judge model. We further propose XParaCon, a novel evaluation metric that quantifies a model's robustness by measuring the standard deviation of accuracies across question variants. Additionally, we implement a reasoning-based, paraphrase-aware Supervised Fine-Tuning (SFT) strategy designed to align models toward semantic invariance. Our experiments demonstrate that this targeted alignment significantly enhances robustness. Notably, fine-tuned lightweight models achieved consistency levels comparable to much larger pre-trained models. These results highlight the efficacy of our approach in mitigating superficial memorization and fostering more robust, reliable LLMs. |
| title | RoParQ: Paraphrase-Aware Alignment of Large Language Models Towards Robustness to Paraphrased Questions |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2511.21568 |