analiza aiqbench
EmoDistill presents a novel offline framework for distilling emotional negotiation skills into smaller language model agents, addressing the vulnerability of overly polite LLMs in adversarial bargaining scenarios. By separating emotion selection via Implicit Q-Learning from expression generation through LoRA-adapted policies, the method demonstrates that explicit emotional channels are critical for maintaining utility across diverse domains. This research advances the intersection of affective computing and reinforcement learning, offering a pathway to build more robust AI agents capable of strategic emotional manipulation without requiring massive compute resources.
EmoDistill: Offline Emotion Skill Distillation for Language Model Agents in Adversarial Negotiation
arXiv:2605.26785v2 Announce Type: replace-cross Abstract: Post-trained LLMs are often optimized to produce helpful, polite, and accommodating responses. In adversarial negotiation, however, such behavior can become a vulnerability: emotionally framed language may influence an agent's bargaining decisions in ways that conflict with its user's objectives. We therefore introduce EmoDistill, an offline framework for distilling emotional negotiation skills from LLM-LLM interactions into smaller language-model agents. Here, an emotional negotiation skill is a state-conditioned behavior that determines which explicit emotion to invoke in a bargaining state and how to realize that emotion as an effective negotiation utterance. EmoDistill learns these two components separately: an Implicit Q-Learning (IQL) selector learns which emotion to express in each bargaining state, while a LoRA-adapted 7B policy learns emotion-conditioned expression through Supervised Fine-Tuning (SFT) and Judge Policy Optimization (JPO). Across four emotion-sensitive negotiation domains, the full EmoDistill policy achieves competitive utility and improves over vanilla and IQL-only baselines in most settings. Emotion-free ablations show that removing the explicit emotion channel substantially reduces overall negotiation utility, while transfer experiments reveal partial, domain-dependent transfer and robustness to unseen LLM counterparties.
Abstract:Post-trained LLMs are often optimized to produce helpful, polite, and accommodating responses. In adversarial negotiation, however, such behavior can become a vulnerability: emotionally framed language may influence an agent's bargaining decisions in ways that conflict with its user's objectives. We therefore introduce EmoDistill, an offline framework for distilling emotional negotiation skills from LLM-LLM interactions into smaller language-model agents. Here, an emotional negotiation skill is a state-conditioned behavior that determines which explicit emotion to invoke in a bargaining state and how to realize that emotion as an effective negotiation utterance. EmoDistill learns these two components separately: an Implicit Q-Learning (IQL) selector learns which emotion to express in each bargaining state, while a LoRA-adapted 7B policy learns emotion-conditioned expression through Supervised Fine-Tuning (SFT) and Judge Policy Optimization (JPO). Across four emotion-sensitive negotiation domains, the full EmoDistill policy achieves competitive utility and improves over vanilla and IQL-only baselines in most settings. Emotion-free ablations show that removing the explicit emotion channel substantially reduces overall negotiation utility, while transfer experiments reveal partial, domain-dependent transfer and robustness to unseen LLM counterparties.
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
arXiv:2605.26785 [cs.CL]
(or arXiv:2605.26785v2 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2605.26785 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 271 cuvinte
Subiecte conexe
What's Continual Learning, and Why Might We Expect To See It In Advanced LLM Agents?
LessWrong · 12 iun. 2026
Biological Superintelligence
LessWrong · 30 iul. 2026
[AINews] Claude Opus 5.5, the new default model for AINews — and everybody cuts prices 40-50%
Latent Space · 23 sept. 2026
[AINews] Meta Muse Wonder & Wonder Full: MSL’s Claude 5.5-tier Frontier Model
Latent Space · 13 oct. 2026