NanoFlux: Adversarial Dual-LLM Evaluation and Distillation for Multi-Domain Reasoning
Raviteja Anantha, Soheil Hor, Teodor Nicola Antoniu, Layne C. Price · 2026
We present NanoFlux, a novel adversarial framework for generating targeted training data to improve LLM reasoning, where adversarially-generated datasets of ≤ 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.Finetuning a 4B-parameter model on NanoFluxgenerated data yields performance gains across diverse domains compared to full-benchmark fine-tuning: +5.9% on mathematical reasoning, +3.6% on scientific reasoning, and +16.6% on medical reasoning, while reducing computational requirements by 3-14×.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, pointing to the value of small, targeted training datasets.