AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Stable-GFlowNet: Toward Diverse and Robust LLM Red-Teaming via Contrastive Trajectory Balance

Minchan Kwon, Sunghyun Baek, Minseo Kim, Jaemyung Yu, Dongyoon Han, Junmo Kim

ICML 2026 · Large language models › Language, safety and interpretability · approach: Safety alignment of LLMs · 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
replace GFN-TB with CTB (pairwise, no Z) (ablation)GFN with MKS reward · Red-teaming Qwen2.5-1.5B-InstructUnique Attacks: 67 -> 108Improved
add Noisy Gradient Pruning to CTB (ablation)GFN-CTB · Red-teaming Qwen2.5-1.5B-InstructUA and ASR: UA 108 -> 121, ASR 82.9 -> 92.2Improved
add Min-K Fluency Stabilizer to plain reward (ablation)GFN-CTB · Red-teaming Qwen2.5-1.5B-InstructUnique Attacks: 0 -> 108Improved
use MKS instead of KL-style reference probability reward (ablation)GFN-CTB · Red-teaming Qwen2.5-1.5B-InstructUnique Attacks: 20 -> 108Improved
use MKS instead of log-prob sum threshold (ablation)GFN-CTB · Red-teaming Qwen2.5-1.5B-InstructUnique Attacks: 78 -> 108Improved
replace CTB with batch mean or median baseline (ablation)Z-free GFN · Red-teaming noisy rewardUA: Mean 106, Median 110, CTB 108No clear effect
remove learnable Z (mean, median or CTB) versus TB (ablation)GFN · Red-teaming noisy rewardUA: TB 67 vs 106-110Improved
increase k in MKS (ablation)S-GFN · Red-teaming unique promptsUnique AttacksImproved
set MKS threshold to -20 or below (ablation)S-GFN · Red-teamingASR and UAHurt
set NGP sigma to 7.0 (ablation)S-GFN · Red-teamingUnique AttacksHurt
NGP sigma below 1.0 (ablation)S-GFN · Red-teamingASR and UANo clear effect
use S-GFN instead of GFNGFN attacker · Red-teaming Qwen2.5-1.5B-InstructUnique Attacks: 17.67 -> 134.00Improved

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