AI Paper Atlas

Research map / Large language models / Instruction and fine-tuning

BARRED: Synthetic Training of Custom Policy Guardrails via Asymmetric Debate

Arnon Mazza, Elad Levi

ICML 2026 · Large language models › Instruction and fine-tuning · approach: Fine-tuning for reasoning data · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
fine-tune a small model on BARRED synthetic data instead of prompting a frontier LLMQwen2.5-14B · four guardrail tasksaccuracyImproved
fine-tune 3B model on BARRED data vs generic guardrailsQwen2.5-3B · four guardrail tasksaccuracyImproved
add debate-based verification vs no verification (ablation)BARRED · Plan verification (human test)accuracy: 0.58 -> 0.85Improved
add debate-based verification vs no verification (ablation)BARRED · Plan verification (synthetic test)accuracy: 0.65 -> 0.99Improved
use multi-agent debate instead of single-agent Self-Refine (ablation)BARRED · Plan verification (human test)accuracy: 0.53 -> 0.85Improved
use Self-Refine instead of no verification (ablation)BARRED · Plan verification (human test)accuracy: 0.58 -> 0.53Hurt
increase number of dimension instantiations (ablation)BARRED · test set coverage and accuracyaccuracy / coverageImproved
scale fine-tuned Qwen2.5 from 1.5B to 14B on simple task (ablation)Qwen2.5 fine-tuned · RepetitionaccuracyNo clear effect
scale fine-tuned Qwen2.5 from 1.5B to 14B on complex tasks (ablation)Qwen2.5 fine-tuned · Privacy and HealthaccuracyImproved
use iterative refinement with judge feedback on rejected samplesBARRED · qualitative examplessample qualityImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy