Research

Paper

AI LLM February 19, 2026

Asking Forever: Universal Activations Behind Turn Amplification in Conversational LLMs

Authors

Zachary Coalson, Bo Fang, Sanghyun Hong

Abstract

Multi-turn interaction length is a dominant factor in the operational costs of conversational LLMs. In this work, we present a new failure mode in conversational LLMs: turn amplification, in which a model consistently prolongs multi-turn interactions without completing the underlying task. We show that an adversary can systematically exploit clarification-seeking behavior$-$commonly encouraged in multi-turn conversation settings$-$to scalably prolong interactions. Moving beyond prompt-level behaviors, we take a mechanistic perspective and identify a query-independent, universal activation subspace associated with clarification-seeking responses. Unlike prior cost-amplification attacks that rely on per-turn prompt optimization, our attack arises from conversational dynamics and persists across prompts and tasks. We show that this mechanism provides a scalable pathway to induce turn amplification: both supply-chain attacks via fine-tuning and runtime attacks through low-level parameter corruptions consistently shift models toward abstract, clarification-seeking behavior across prompts. Across multiple instruction-tuned LLMs and benchmarks, our attack substantially increases turn count while remaining compliant. We also show that existing defenses offer limited protection against this emerging class of failures.

Metadata

arXiv ID: 2602.17778
Provider: ARXIV
Primary Category: cs.LG
Published: 2026-02-19
Fetched: 2026-02-23 05:33

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2602.17778v1</id>\n    <title>Asking Forever: Universal Activations Behind Turn Amplification in Conversational LLMs</title>\n    <updated>2026-02-19T19:21:09Z</updated>\n    <link href='https://arxiv.org/abs/2602.17778v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2602.17778v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Multi-turn interaction length is a dominant factor in the operational costs of conversational LLMs. In this work, we present a new failure mode in conversational LLMs: turn amplification, in which a model consistently prolongs multi-turn interactions without completing the underlying task. We show that an adversary can systematically exploit clarification-seeking behavior$-$commonly encouraged in multi-turn conversation settings$-$to scalably prolong interactions. Moving beyond prompt-level behaviors, we take a mechanistic perspective and identify a query-independent, universal activation subspace associated with clarification-seeking responses. Unlike prior cost-amplification attacks that rely on per-turn prompt optimization, our attack arises from conversational dynamics and persists across prompts and tasks. We show that this mechanism provides a scalable pathway to induce turn amplification: both supply-chain attacks via fine-tuning and runtime attacks through low-level parameter corruptions consistently shift models toward abstract, clarification-seeking behavior across prompts. Across multiple instruction-tuned LLMs and benchmarks, our attack substantially increases turn count while remaining compliant. We also show that existing defenses offer limited protection against this emerging class of failures.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.LG'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CR'/>\n    <published>2026-02-19T19:21:09Z</published>\n    <arxiv:comment>Pre-print</arxiv:comment>\n    <arxiv:primary_category term='cs.LG'/>\n    <author>\n      <name>Zachary Coalson</name>\n    </author>\n    <author>\n      <name>Bo Fang</name>\n    </author>\n    <author>\n      <name>Sanghyun Hong</name>\n    </author>\n  </entry>"
}