NanoFlux: Adversarial Dual-LLM Evaluation and Distillation For Multi-Domain Reasoning

• URL: http://arxiv.org/abs/2509.23252v2
• Date: Sat, 04 Oct 2025 01:19:03 GMT
• Title: NanoFlux: Adversarial Dual-LLM Evaluation and Distillation For Multi-Domain Reasoning
• Abstract summary: NanoFlux is a novel adversarial framework for generating targeted training data to improve LLM reasoning.<n>The framework employs a competitive dynamic between models alternating as Attacker and Defender, supervised by a tool-augmented Judge.<n>Fine-tuning a 4B- parameter model on NanoFlux-generated data yields performance gains across diverse domains compared to full-benchmark fine-tuning.
• Score: 5.461464418720756
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We present NanoFlux, a novel adversarial framework for generating targeted training data to improve LLM reasoning, where adversarially-generated datasets containing fewer than 200 examples outperform conventional fine-tuning approaches. The framework employs a competitive dynamic between models alternating as Attacker and Defender, supervised by a tool-augmented Judge, synthesizing multi-step questions with explanatory annotations that target specific reasoning capabilities. Fine-tuning a 4B-parameter model on NanoFlux-generated data yields performance gains across diverse domains compared to full-benchmark fine-tuning: +5.9% on mathematical reasoning (GSMHard), +3.6% on scientific reasoning (GenomeBench), and +16.6% on medical reasoning (MultiMedQA), while reducing computational requirements by 3-14x. Ablation studies reveal a non-monotonic relationship between dataset characteristics and model performance, uncovering domain-specific optimal points for question complexity and reasoning quality. NanoFlux automates training data generation through embedding-based novelty filtering, tool-augmented evaluation, and multi-hop reasoning, suggesting that future model improvements may lie in the intelligent synthesis of small, precisely targeted training datasets.

Related papers

• Dual-Phase LLM Reasoning: Self-Evolved Mathematical Frameworks [48.105258051884384]
  This paper proposes a new two-stage training framework that enhances models' self-correction capabilities.<n>During the first stage, a multi-turn dialogue strategy guides the model to generate long chain-of-thought (CoT) data.<n>The second stage employs a difficulty-aware rejection sampling mechanism to dynamically optimize data distribution.
  arXiv  Detail & Related papers  (2026-01-09T08:19:11Z)
• Enhanced Data-Driven Product Development via Gradient Based Optimization and Conformalized Monte Carlo Dropout Uncertainty Estimation [1.683673215692404]
  Data-Driven Product Development (DDPD) leverages data to learn the relationship between product design specifications and resulting properties.<n>We train a neural network on past experiments and apply Projected Gradient Descent to identify optimal input features that maximize performance.<n>We integrate uncertainty estimation via emphConformalised Monte Carlo Dropout (ConfMC) to provide model-agnostic, finite-sample coverage guarantees under data exchangeability.
  arXiv  Detail & Related papers  (2026-01-02T07:55:02Z)
• One Model to Critique Them All: Rewarding Agentic Tool-Use via Efficient Reasoning [54.580646706013965]
  Reward models (RMs) play a critical role in aligning large language models with human preferences.<n>We introduce ToolRM, a family of lightweight generative RMs tailored for general tool-use scenarios.<n>To build these models, we propose a novel pipeline that constructs pairwise preference data using rule-based scoring and multidimensional sampling.
  arXiv  Detail & Related papers  (2025-10-30T06:08:27Z)
• Hierarchical Evaluation Function: A Multi-Metric Approach for Optimizing Demand Forecasting Models [0.479839492673697]
  The Hierarchical Evaluation Function (HEF) is proposed as a multi-metric framework for hyperparameter optimization.<n>HEF integrates explanatory power (R2), sensitivity to extreme errors (RMSE), and average accuracy (MAE)<n>The performance of HEF was assessed using four widely recognized benchmark datasets in the forecasting domain.
  arXiv  Detail & Related papers  (2025-08-18T16:25:49Z)
• A Comprehensive Taxonomy of Negation for NLP and Neural Retrievers [57.57320129313161]
  We introduce a taxonomy of negation that derives from philosophical, linguistic, and logical definitions.<n>We generate two benchmark datasets that can be used to evaluate the performance of neural information retrieval models.<n>We propose a logic-based classification mechanism that can be used to analyze the performance of retrieval models on existing datasets.
  arXiv  Detail & Related papers  (2025-07-30T02:44:20Z)
• Model-agnostic Mitigation Strategies of Data Imbalance for Regression [0.0]
  Data imbalance persists as a pervasive challenge in regression tasks, introducing bias in model performance and undermining predictive reliability.<n>We present advanced mitigation techniques, which build upon and improve existing sampling methods.<n>We demonstrate that constructing an ensemble of models -- one trained with imbalance mitigation and another without -- can significantly reduce these negative effects.
  arXiv  Detail & Related papers  (2025-06-02T09:46:08Z)
• MindGYM: What Matters in Question Synthesis for Thinking-Centric Fine-Tuning? [51.85759493254735]
  MindGYM is a structured and scalable framework for question synthesis.<n>It infuses high-level reasoning objectives to shape the model's synthesis behavior.<n>It composes more complex multi-hop questions based on QA seeds for deeper reasoning.
  arXiv  Detail & Related papers  (2025-03-12T16:03:03Z)
• On Machine Learning Approaches for Protein-Ligand Binding Affinity Prediction [2.874893537471256]
  This study evaluates the performance of classical tree-based models and advanced neural networks in protein-ligand binding affinity prediction. We show that combining 2D and 3D model strengths improves active learning outcomes beyond current state-of-the-art approaches.
  arXiv  Detail & Related papers  (2024-07-15T13:06:00Z)
• GenBench: A Benchmarking Suite for Systematic Evaluation of Genomic Foundation Models [56.63218531256961]
  We introduce GenBench, a benchmarking suite specifically tailored for evaluating the efficacy of Genomic Foundation Models. GenBench offers a modular and expandable framework that encapsulates a variety of state-of-the-art methodologies. We provide a nuanced analysis of the interplay between model architecture and dataset characteristics on task-specific performance.
  arXiv  Detail & Related papers  (2024-06-01T08:01:05Z)
• When predict can also explain: few-shot prediction to select better neural latents [2.8933605229876664]
  One widely-used method, $textitco-smoothing$, involves jointly estimating latent variables and predicting observations along held-out channels.<n>We show that models with high co-smoothing can have arbitrary extraneous dynamics in their latent representations.<n>By cross-decoding the latent variables of all model pairs with high co-smoothing, we identify models with minimal extraneous dynamics.
  arXiv  Detail & Related papers  (2024-05-23T10:48:30Z)
• Discover, Explanation, Improvement: An Automatic Slice Detection Framework for Natural Language Processing [72.14557106085284]
  slice detection models (SDM) automatically identify underperforming groups of datapoints. This paper proposes a benchmark named "Discover, Explain, improve (DEIM)" for classification NLP tasks. Our evaluation shows that Edisa can accurately select error-prone datapoints with informative semantic features.
  arXiv  Detail & Related papers  (2022-11-08T19:00:00Z)
• Quaternion Factorization Machines: A Lightweight Solution to Intricate Feature Interaction Modelling [76.89779231460193]
  factorization machine (FM) is capable of automatically learning high-order interactions among features to make predictions without the need for manual feature engineering. We propose the quaternion factorization machine (QFM) and quaternion neural factorization machine (QNFM) for sparse predictive analytics.
  arXiv  Detail & Related papers  (2021-04-05T00:02:36Z)

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.