Peer-Predictive Self-Training for Language Model Reasoning

• URL: http://arxiv.org/abs/2604.13356v1
• Date: Tue, 14 Apr 2026 23:29:44 GMT
• Title: Peer-Predictive Self-Training for Language Model Reasoning
• Abstract summary: We propose a label-free fine-tuning framework in which multiple language models improve collaboratively.<n>Given a prompt question, the models generate responses sequentially; the final aggregated answer serves as an internal target for learning.<n>On mathematical reasoning benchmarks, PST improves exact-match accuracy by 2.2 to 4.3 percentage points across Gemma-2-2B, LLaMA-3.2-1B, and Qwen-2.5-1.5B.
• Score: 15.499995918997756
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Mechanisms for continued self-improvement of language models without external supervision remain an open challenge. We propose Peer-Predictive Self-Training (PST), a label-free fine-tuning framework in which multiple language models improve collaboratively by leveraging a cross-model aggregated response as an internal training signal. Given a prompt question, the models generate responses sequentially; the final aggregated answer, often more reliable than individual responses in practice, serves as an internal target for learning. We measure how informative each intermediate response is about the aggregate using pointwise mutual information (PMI), and use this signal to scale self-training updates. Responses already aligned with the aggregate are updated less, while less informative or misaligned responses are updated more. On mathematical reasoning benchmarks (SimulEq, Math500, and MultiArith), PST improves exact-match accuracy by 2.2 to 4.3 percentage points across Gemma-2-2B, LLaMA-3.2-1B, and Qwen-2.5-1.5B, and reduces the average generator-verifier gap (GV-Gap) by 26 to 40 percent, while requiring no external supervision or teacher-student hierarchy and relying solely on cross-model interactions. These results suggest that cross-model generations and peer-predictive feedback can serve as an effective approach for self-supervised training.

Related papers

• NeoHorse-1: Towards Recursive Self-Improvement via Agentic Post-Training with Routing Harness [76.6324710887112]
  NeoHorse-1 is a family of agent-native models developed to explore this path through agentic post-training.<n>Our system combines a heterogeneous model pool with intelligent routing, recording the predicted capability demand, selected service tier, and subsequent interaction for each user turn.
  arXiv  Detail & Related papers  (2026-09-08T03:14:54Z)
• Ask, Solve, Generate: Self-Evolving Unified Multimodal Understanding and Generation via Self-Consistency Rewards [52.42920996842378]
  Most unified large multimodal models (LMMs) that support both visual understanding and image generation rely on curated post-training supervision.<n>We propose a self-evolving training framework with three internal roles: a Proposer that generates visual questions, a solver that answers and evaluates them, and a Generator that synthesizes images.<n>For image generation, we design a multi-scale internal evaluation scheme that combines question-answer fidelity scoring with cycle-consistent captioning.
  arXiv  Detail & Related papers  (2026-06-25T17:59:57Z)
• Querit-Reranker: Training Compact Multilingual Rerankers via Efficient Label-Free Distribution Adaptation [43.66118282567413]
  We present Querit-Reranker, a family of multilingual cross-encoder rerankers trained with a data-centric pipeline.<n>Our pipeline first learns general relevance modeling from large-scale ranking-oriented data, then adapts to target distributions.<n>Querit-Reranker-A0.4B improves average nDCG@10 from 54.11 to 59.28 on BEIR and from 59.87 to 67.70 on MIRACL.
  arXiv  Detail & Related papers  (2026-06-17T13:06:47Z)
• When Models Judge Themselves: Unsupervised Self-Evolution for Multimodal Reasoning [30.162034423611292]
  We propose an unsupervised self-evolution training framework for multimodal reasoning.<n>It achieves stable performance improvements without using human-annotated answers or external reward models.<n>Our method consistently improves reasoning performance and generalization on five mathematical reasoning benchmarks.
  arXiv  Detail & Related papers  (2026-03-22T15:22:19Z)
• Summarize Before You Speak with ARACH: A Training-Free Inference-Time Plug-In for Enhancing LLMs via Global Attention Reallocation [9.508727214134106]
  Large language models (LLMs) achieve remarkable performance, yet further gains often require costly training.<n>This has motivated growing interest in post-training techniques-especially training-free approaches that improve models at inference time without updating weights.<n>We propose ARACH(Attention Reallocation via an Adaptive Context Hub), a training-free inference-time plug-in that augments LLMs with an adaptive context hub to aggregate context and reallocate attention.
  arXiv  Detail & Related papers  (2026-03-10T06:07:54Z)
• Self-Consistency as a Free Lunch: Reducing Hallucinations in Vision-Language Models via Self-Reflection [71.8243083897721]
  Vision-language models often hallucinate details, generating non-existent objects or inaccurate attributes that compromise output reliability.<n>We present a novel framework that leverages the model's self-consistency between long responses and short answers to generate preference pairs for training.
  arXiv  Detail & Related papers  (2025-09-27T10:37:11Z)
• SPaR: Self-Play with Tree-Search Refinement to Improve Instruction-Following in Large Language Models [88.29990536278167]
  We introduce SPaR, a self-play framework integrating tree-search self-refinement to yield valid and comparable preference pairs.<n>Our experiments show that a LLaMA3-8B model, trained over three iterations guided by SPaR, surpasses GPT-4-Turbo on the IFEval benchmark without losing general capabilities.
  arXiv  Detail & Related papers  (2024-12-16T09:47:43Z)
• Self-Judge: Selective Instruction Following with Alignment Self-Evaluation [27.69410513313001]
  We study the study of selective instruction following, whereby the system declines to execute instructions if the anticipated response quality is low. We introduce Self-J, a novel self-training framework for developing judge models without needing human-annotated quality scores.
  arXiv  Detail & Related papers  (2024-09-02T04:14:13Z)
• Unifying Language Learning Paradigms [96.35981503087567]
  We present a unified framework for pre-training models that are universally effective across datasets and setups. We show how different pre-training objectives can be cast as one another and how interpolating between different objectives can be effective. Our model also achieve strong results at in-context learning, outperforming 175B GPT-3 on zero-shot SuperGLUE and tripling the performance of T5-XXL on one-shot summarization.
  arXiv  Detail & Related papers  (2022-05-10T19:32:20Z)
• Momentum Pseudo-Labeling for Semi-Supervised Speech Recognition [55.362258027878966]
  We present momentum pseudo-labeling (MPL) as a simple yet effective strategy for semi-supervised speech recognition. MPL consists of a pair of online and offline models that interact and learn from each other, inspired by the mean teacher method. The experimental results demonstrate that MPL effectively improves over the base model and is scalable to different semi-supervised scenarios.
  arXiv  Detail & Related papers  (2021-06-16T16:24:55Z)
• AutoFIS: Automatic Feature Interaction Selection in Factorization Models for Click-Through Rate Prediction [75.16836697734995]
  We propose a two-stage algorithm called Automatic Feature Interaction Selection (AutoFIS) AutoFIS can automatically identify important feature interactions for factorization models with computational cost just equivalent to training the target model to convergence. AutoFIS has been deployed onto the training platform of Huawei App Store recommendation service.
  arXiv  Detail & Related papers  (2020-03-25T06:53:54Z)

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.