# Calling the Bluff: Detecting Ever-Shifting Harmful Chat Dialogue via Ordered Reasoning Chain Regularization

arXiv: https://arxiv.org/abs/2608.08451 · HTML: https://mlsanity.com/paper/2608.08451/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning.
- Builds on 9 of its 38 references that are in the map.
- First posted on arXiv 9 Aug 2026.

## Placed in

- LLM reasoning / [Reasoning benchmarks and evaluation](https://mlsanity.com/lr/reasoning-benchmarks-and-evaluation/)

## Builds on (most cited first, 9 of 9)

- [LMSYS-Chat-1M: A Large-Scale Real-World LLM Conversation Dataset](https://mlsanity.com/paper/2309.11998/) (2023)
- [A Mechanistic Understanding of Alignment Algorithms: A Case Study on DPO and Toxicity](https://mlsanity.com/paper/2401.01967/) (2024)
- [Safety Assessment of Chinese Large Language Models](https://mlsanity.com/paper/2304.10436/) (2023)
- [ReasoningShield: Safety Detection over Reasoning Traces of Large Reasoning Models](https://mlsanity.com/paper/2505.17244/) (2025)
- [GW-MoE: Resolving Uncertainty in MoE Router with Global Workspace Theory](https://mlsanity.com/paper/2406.12375/) (2024)
- [ChineseSafe: A Chinese Benchmark for Evaluating Safety in Large Language Models](https://mlsanity.com/paper/2410.18491/) (2024)
- [ChineseHarm-Bench: A Chinese Harmful Content Detection Benchmark](https://mlsanity.com/paper/2506.10960/) (2025)
- [JADE: A Linguistics-based Safety Evaluation Platform for Large Language Models](https://mlsanity.com/paper/2311.00286/) (2023)
- [Toxicity in Online Platforms and AI Systems: A Survey of Needs, Challenges, Mitigations, and Future Directions](https://mlsanity.com/paper/2509.25539/) (2025)

## Abstract (arXiv)

Harmful chat dialogues are ever-shifting through type-shifting and lexical evasion, yet we find they share invariant principles, i.e., an Ordered Reasoning Chain (ORC) of recurring topics, harm language indicators, severity hierarchies, and type characteristics, which can help us capture the key information in the frequently changing lexical expressions. We propose BRACE, which encodes the ORC as four differentiable stages (Topic -> Indicator -> Severity -> Type) with intermediate supervision, serving as a structured regularizer blended with direct heads, and supported by prototype-based feature augmentation and feature path disentanglement. The evaluation results show that, across 4 domains and 5 harm categories, BRACE achieves harm-type macro F1 of 0.934 (RoBERTa-wwm-ext, 3-seed mean), with decoder backbones (Qwen3-1.7B LoRA) reaching 0.949. Ablation studies show that all components contribute to BRACE, and the structural decomposition of ORC enables BRACE to distinguish harmful types with semantic ambiguity. Disclaimer: This paper may contain content that is disturbing to some readers.