Paper Detail
Xiachong Feng, Yi Jiang, Xiaocheng Feng, Deyi Yin, Libo Qin, Yangfan Ye, Lei Huang, Weitao Ma, Yuxuan Gu, Chonghan Qin, Bing Qin, Lingpeng Kong
Social intelligence, the ability to navigate complex interpersonal interactions, presents a fundamental challenge for language agents. Training such agents via reinforcement learning requires solving the credit assignment problem: determining how individual utterances contribute to multi-turn dialogue outcomes. Existing approaches directly employ language models to distribute episode-level rewards, yielding attributions that are retrospective and lack theoretical grounding. We propose SAVOIR (ShApley Value fOr SocIal RL), a novel principled framework grounded in cooperative game theory. Our approach combines two complementary principles: expected utility shifts evaluation from retrospective attribution to prospective valuation, capturing an utterance's strategic potential for enabling favorable future trajectories; Shapley values ensure fair credit distribution with axiomatic guarantees of efficiency, symmetry, and marginality. Experiments on the SOTOPIA benchmark demonstrate that SAVOIR achieves new state-of-the-art performance across all evaluation settings, with our 7B model matching or exceeding proprietary models including GPT-4o and Claude-3.5-Sonnet. Notably, even large reasoning models consistently underperform, suggesting social intelligence requires qualitatively different capabilities than analytical reasoning.
No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.
No ranking explanation is available yet.
No tags.
@misc{feng2026savoir,
title = {SAVOIR: Learning Social Savoir-Faire via Shapley-based Reward Attribution},
author = {Xiachong Feng and Yi Jiang and Xiaocheng Feng and Deyi Yin and Libo Qin and Yangfan Ye and Lei Huang and Weitao Ma and Yuxuan Gu and Chonghan Qin and Bing Qin and Lingpeng Kong},
year = {2026},
abstract = {Social intelligence, the ability to navigate complex interpersonal interactions, presents a fundamental challenge for language agents. Training such agents via reinforcement learning requires solving the credit assignment problem: determining how individual utterances contribute to multi-turn dialogue outcomes. Existing approaches directly employ language models to distribute episode-level rewards, yielding attributions that are retrospective and lack theoretical grounding. We propose SAVOIR (Sh},
url = {https://huggingface.co/papers/2604.18982},
keywords = {reinforcement learning, credit assignment problem, language models, dialogue outcomes, cooperative game theory, expected utility shifts, Shapley values, social intelligence, language agents, SOTOPIA benchmark, episode-level rewards, code available, huggingface daily},
eprint = {2604.18982},
archiveprefix = {arXiv},
}
{}