Gained in Translation: Privileged Pairwise Judges Enhance Multilingual Reasoning

• URL: http://arxiv.org/abs/2601.18722v1
• Date: Mon, 26 Jan 2026 17:46:44 GMT
• Title: Gained in Translation: Privileged Pairwise Judges Enhance Multilingual Reasoning
• Abstract summary: We introduce textttSP3F, a framework for enhancing multilingual reasoning without textitany data in the target language(s)<n>First, we supervise fine-tune (SFT) on translated versions of English question-answer pairs to raise base model correctness.<n>Second, we perform RL with feedback from a pairwise judge in a self-play fashion, with the judge receiving the English reference response as textitprivileged information.
• Score: 74.00326879493225
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: When asked a question in a language less seen in its training data, current reasoning large language models (RLMs) often exhibit dramatically lower performance than when asked the same question in English. In response, we introduce \texttt{SP3F} (Self-Play with Privileged Pairwise Feedback), a two-stage framework for enhancing multilingual reasoning without \textit{any} data in the target language(s). First, we supervise fine-tune (SFT) on translated versions of English question-answer pairs to raise base model correctness. Second, we perform RL with feedback from a pairwise judge in a self-play fashion, with the judge receiving the English reference response as \textit{privileged information}. Thus, even when none of the model's responses are completely correct, the privileged pairwise judge can still tell which response is better. End-to-end, \texttt{SP3F} greatly improves base model performance, even outperforming fully post-trained models on multiple math and non-math tasks with less than of the training data across the single-language, multilingual, and generalization to unseen language settings.

Related papers

• Choosing the Right Language Mode at Inference Time for Multilingual Reliability [3.6095388702618405]
  Large language models often struggle to reason in low- to mid-resource languages.<n>Prior work has shown that translation can improve multilingual reasoning by helping models access stronger English-centric representations.<n>This raises a central question: How much translation is needed for multilingual large language models to reason reliably.
  arXiv  Detail & Related papers  (2026-09-04T02:36:26Z)
• Align to the Pivot: Dual Alignment with Self-Feedback for Multilingual Math Reasoning [71.4175109189942]
  We present Pivot-Aligned Self-Feedback Multilingual Reasoning (PASMR)<n>This approach designates the model's primary language as the pivot language.<n>It establishes a cross-lingual self-feedback mechanism without relying on external correct answers or reward models.
  arXiv  Detail & Related papers  (2026-01-25T03:20:00Z)
• The Power of Question Translation Training in Multilingual Reasoning: Broadened Scope and Deepened Insights [108.40766216456413]
  We propose a question alignment framework to bridge the gap between large language models' English and non-English performance. Experiment results show it can boost multilingual performance across diverse reasoning scenarios, model families, and sizes. We analyze representation space, generated response and data scales, and reveal how question translation training strengthens language alignment within LLMs.
  arXiv  Detail & Related papers  (2024-05-02T14:49:50Z)
• Question Translation Training for Better Multilingual Reasoning [108.10066378240879]
  Large language models show compelling performance on reasoning tasks but they tend to perform much worse in languages other than English. A typical solution is to translate instruction data into all languages of interest, and then train on the resulting multilingual data, which is called translate-training. In this paper we explore the benefits of question alignment, where we train the model to translate reasoning questions into English by finetuning on X-English parallel question data.
  arXiv  Detail & Related papers  (2024-01-15T16:39:10Z)
• Ask Again, Then Fail: Large Language Models' Vacillations in Judgment [28.74246375289661]
  We observe that current conversational language models often waver in their judgments when faced with follow-up questions. We introduce a textscFollow-up Questioning Mechanism along with two metrics to quantify this inconsistency. We develop a training-based framework textscUnwavering-FQ that teaches language models to maintain their originally correct judgments.
  arXiv  Detail & Related papers  (2023-10-03T16:08:41Z)
• Making Retrieval-Augmented Language Models Robust to Irrelevant Context [55.564789967211844]
  An important desideratum of RALMs, is that retrieved information helps model performance when it is relevant. Recent work has shown that retrieval augmentation can sometimes have a negative effect on performance.
  arXiv  Detail & Related papers  (2023-10-02T18:52:35Z)
• MuCoT: Multilingual Contrastive Training for Question-Answering in Low-resource Languages [4.433842217026879]
  Multi-lingual BERT-based models (mBERT) are often used to transfer knowledge from high-resource languages to low-resource languages. We augment the QA samples of the target language using translation and transliteration into other languages and use the augmented data to fine-tune an mBERT-based QA model. Experiments on the Google ChAII dataset show that fine-tuning the mBERT model with translations from the same language family boosts the question-answering performance.
  arXiv  Detail & Related papers  (2022-04-12T13:52:54Z)
• Switch Point biased Self-Training: Re-purposing Pretrained Models for Code-Switching [44.034300203700234]
  Code-switching is a ubiquitous phenomenon due to the ease of communication it offers in multilingual communities. We propose a self training method to repurpose the existing pretrained models using a switch-point bias. Our approach performs well on both tasks by reducing the gap between the switch point performance.
  arXiv  Detail & Related papers  (2021-11-01T19:42:08Z)
• Multilingual Answer Sentence Reranking via Automatically Translated Data [97.98885151955467]
  We present a study on the design of multilingual Answer Sentence Selection (AS2) models, which are a core component of modern Question Answering (QA) systems. The main idea is to transfer data, created from one resource rich language, e.g., English, to other languages, less rich in terms of resources.
  arXiv  Detail & Related papers  (2021-02-20T03:52:08Z)
• Cross-lingual Spoken Language Understanding with Regularized Representation Alignment [71.53159402053392]
  We propose a regularization approach to align word-level and sentence-level representations across languages without any external resource. Experiments on the cross-lingual spoken language understanding task show that our model outperforms current state-of-the-art methods in both few-shot and zero-shot scenarios.
  arXiv  Detail & Related papers  (2020-09-30T08:56:53Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.