# Do You Feel Comfortable? Detecting Hidden Conversational Escalation in AI Chatbots

arXiv: https://arxiv.org/abs/2512.06193 · HTML: https://mlsanity.com/paper/2512.06193/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 10 of its 27 references that are in the map.
- First posted on arXiv 5 Dec 2025.

## Builds on (most cited first, 10 of 10)

- [The Internal State of an LLM Knows When It's Lying](https://mlsanity.com/paper/2304.13734/) (2023)
- [HateBERT: Retraining BERT for Abusive Language Detection in English](https://mlsanity.com/paper/2010.12472/) (2020)
- [Survey on Evaluation Methods for Dialogue Systems](https://mlsanity.com/paper/1905.04071/) (2019)
- [On the Safety of Conversational Models: Taxonomy, Dataset, and Benchmark](https://mlsanity.com/paper/2110.08466/) (2021)
- [The Scales of Justitia: A Comprehensive Survey on Safety Evaluation of LLMs](https://mlsanity.com/paper/2506.11094/) (2025)
- [LLM-Microscope: Uncovering the Hidden Role of Punctuation in Context Memory of Transformers](https://mlsanity.com/paper/2502.15007/) (2025)
- [MinorBench: A hand-built benchmark for content-based risks for children](https://mlsanity.com/paper/2503.10242/) (2025)
- [DeMod: A Holistic Tool with Explainable Detection and Personalized Modification for Toxicity Censorship](https://mlsanity.com/paper/2411.01844/) (2024)
- [Revealing Hidden Mechanisms of Cross-Country Content Moderation with Natural Language Processing](https://mlsanity.com/paper/2503.05280/) (2025)
- [Towards Generalized Offensive Language Identification](https://mlsanity.com/paper/2407.18738/) (2024)

## Built on by (most cited first, 1 of 1)

- [Understanding Risk and Dependency in AI Chatbot Use from User Discourse](https://mlsanity.com/paper/2602.09339/) (2026)

## Abstract (arXiv)

Large Language Models (LLM) are increasingly integrated into everyday interactions, serving not only as information assistants but also as emotional companions. Even in the absence of explicit toxicity, repeated emotional reinforcement or affective drift can gradually escalate distress in a form of textitimplicit harm that traditional toxicity filters fail to detect. Existing guardrail mechanisms often rely on external classifiers or clinical rubrics that may lag behind the nuanced, real-time dynamics of a developing conversation. To address this gap, we propose GAUGE (Guarding Affective Utterance Generation Escalation), logit-based framework for the real-time detection of hidden conversational escalation. GAUGE measures how an LLM's output probabilistically shifts the affective state of a dialogue.