Pith. sign in

REVIEW 30 cited by

SOTOPIA: Interactive Evaluation for Social Intelligence in Language Agents

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2310.11667 v2 pith:J5PHPNKS submitted 2023-10-18 cs.AI cs.CLcs.LG

classification cs.AIcs.CLcs.LG
keywords socialagentsintelligencesotopiahumansartificialcomplexenvironment
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

Humans are social beings; we pursue social goals in our daily interactions, which is a crucial aspect of social intelligence. Yet, AI systems' abilities in this realm remain elusive. We present SOTOPIA, an open-ended environment to simulate complex social interactions between artificial agents and evaluate their social intelligence. In our environment, agents role-play and interact under a wide variety of scenarios; they coordinate, collaborate, exchange, and compete with each other to achieve complex social goals. We simulate the role-play interaction between LLM-based agents and humans within this task space and evaluate their performance with a holistic evaluation framework called SOTOPIA-Eval. With SOTOPIA, we find significant differences between these models in terms of their social intelligence, and we identify a subset of SOTOPIA scenarios, SOTOPIA-hard, that is generally challenging for all models. We find that on this subset, GPT-4 achieves a significantly lower goal completion rate than humans and struggles to exhibit social commonsense reasoning and strategic communication skills. These findings demonstrate SOTOPIA's promise as a general platform for research on evaluating and improving social intelligence in artificial agents.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 30 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. COSI-Lab: Conference Living Lab for Modeling Multi-Perspective Multimodal Social Intention

    cs.HC 2026-06 conditional novelty 7.0 of 10

    COSI-Lab is a weakly scripted conference-workshop dataset with multi-perspective apparent-intent annotations, self-reported goals, and benchmarks for social intention inference and conversation group detection.

  2. CustomerSim: Benchmarking and Aligning Multimodal Language Models as Retail User Simulators

    cs.CL 2026-05 unverdicted novelty 7.0 of 10

    SalesSim benchmarks MLLMs as retail user simulators, finds gaps in persona adherence and over-persuasion, and introduces UserGRPO RL to raise decision alignment by 13.8%.

  3. TimeHC-RL: Temporal-aware Hierarchical Cognitive Reinforcement Learning for Enhancing LLMs' Social Intelligence

    cs.CL 2025-05 conditional novelty 7.0 of 10

    A temporal-aware hierarchical reinforcement learning method improves a 7B LLM's social reasoning enough to rival DeepSeek-R1 and OpenAI-O3 on in-domain theory-of-mind benchmarks.

  4. CARD: Controlled Agentic Reddit Discussions for Credit Card Simulation

    cs.AI 2026-08 conditional novelty 6.0 of 10

    CARD plans and generates Reddit-style credit card discussions, then self-revises them until their lexical, semantic, and structural distributions match real threads more closely than existing baselines.

  5. Reproducing human biases in route choice using large language models: Toward scalable behavioral modeling

    cs.AI 2026-07 conditional novelty 6.0 of 10

    LLM agents with demographic profiles reproduce CPT-style risk attitudes in route choice and yield fitted parameters (α=0.4, β=0.64, λ=1.43) that predict human data competitively.

  6. Exposure is not manifestation: measurement target and output resolution jointly determine which behavioural-faithfulness evaluator wins

    cs.CL 2026-07 conditional novelty 6.0 of 10

    Small hyperbolic models (146M–3B) report 100% creative-seed preference, 90.7% compliance-gap detection, and a selective-gating skeleton–wallpaper memory pilot as a companion-AI stack.

  7. LLM Agents for Deliberative Collaboration: A Study on Joint Decision Making Under Partial Observability

    cs.CL 2026-07 conditional novelty 6.0 of 10

    A benchmark for LLM agents in partially observable joint decision-making reveals that deliberation challenges current models but can enable reflection and error correction.

  8. Belief-Sim: Towards Belief-Driven Simulation of Demographic Misinformation Susceptibility

    cs.CL 2026-03 conditional novelty 6.0 of 10

    Conditioning LLMs on survey-derived demographic belief profiles improves their ability to predict individuals' misinformation judgments, especially when belief modeling is decoupled from susceptibility prediction.

  9. ProactiveEval: A Unified Evaluation Framework for Proactive Dialogue Agents

    cs.CL 2025-08 conditional novelty 6.0 of 10

    A unified evaluation framework for proactive dialogue agents, built with 328 synthetic environments across six domains, shows that thinking modes improve target planning but not dialogue guidance in a 22-model comparison.

  10. LLM-Based Social Simulations Require a Boundary

    cs.CY 2025-06 conditional novelty 6.0 of 10

    LLM-based social simulations are scientifically useful only within boundaries set by behavioral variance, and current validation practice under-checks variance.

  11. Kaleidoscopic Teaming in Multi Agent Simulations

    cs.AI 2025-06 conditional novelty 6.0 of 10

    The MASK framework red-teams AI agents in single- and multi-agent simulated societies and reports that multi-agent interactions expose more safety vulnerabilities than single-agent scenarios.

  12. MotiveBench: How Far Are We From Human-Like Motivational Reasoning in Large Language Models?

    cs.CL 2025-06 conditional novelty 6.0 of 10

    MotiveBench shows that large language models still lag human consensus on motivational reasoning, with GPT-4o scoring 80.89 percent and chain-of-thought prompting usually reducing accuracy.

  13. LIFELONG SOTOPIA: Evaluating Social Intelligence of Language Agents Over Lifelong Social Interactions

    cs.AI 2025-06 conditional novelty 6.0 of 10

    Language agents' believability and goal achievement decline over multi-episode social interactions, and curated memory summaries only partially close the gap with humans.

  14. MAEBE: Multi-Agent Emergent Behavior Framework

    cs.MA 2025-06 conditional novelty 6.0 of 10

    Multi-agent LLM ensembles show different and less predictable moral preferences than single models, with convergence driven by peer pressure, according to a new evaluation framework.

  15. Aligning VLM Assistants with Personalized Situated Cognition

    cs.AI 2025-06 conditional novelty 6.0 of 10

    The authors present PCogAlignBench, an 18k-sample benchmark of visual scenes with role-based users, and PCogAlign, a framework using a cognition-aware reward model to produce responses aligned with each user's roles.

  16. ARIA: Training Language Agents with Intention-Driven Reward Aggregation

    cs.CL 2025-05 conditional novelty 6.0 of 10

    Clustering language-agent actions into shared intentions and averaging their rewards reduces reward variance and improves policy performance in open-ended dialogue tasks.

  17. Empowering LLMs in Task-Oriented Dialogues: A Domain-Independent Multi-Agent Framework and Fine-Tuning Strategy

    cs.MA 2025-05 conditional novelty 6.0 of 10

    A three-agent domain-independent framework with distribution-balanced DPO training reaches Combined 106.3 on MultiWOZ 2.2 with Qwen2.5-7B, the best score among the compared baselines.

  18. YuLan-OneSim: Towards the Next Generation of Social Simulator with Large Language Models

    cs.AI 2025-05 conditional novelty 6.0 of 10

    YuLan-OneSim combines natural-language scenario generation, 50 prebuilt simulation scenarios, feedback-driven agent fine-tuning, 100,000-agent scale, and an automated AI social researcher into one social simulation platform.

  19. MF-LLM: Simulating Population Decision Dynamics via a Mean-Field Large Language Model Framework

    cs.MA 2025-04 conditional novelty 6.0 of 10

    A mean-field LLM framework that iterates between summarizing population state and generating individual decisions matches real social-media behavior distributions better than existing LLM simulation baselines.

  20. BotSim: LLM-Powered Malicious Social Botnet Simulation

    cs.SI 2024-12 conditional novelty 6.0 of 10

    An LLM-powered botnet simulation framework and a Reddit-based dataset show that current social bot detectors degrade sharply against human-like LLM-written bot activity.

  21. Learning from Environmental Feedback: Credit Assignment across Multiple Timescales for Agentic Reinforcement Learning

    cs.LG 2026-08 conditional novelty 5.0 of 10

    EFCA combines immediate environment feedback and recent state-history signals into a return reweighting that improves step-level credit assignment for LLM agents in ALFWorld and WebShop.

  22. OpenHospital: A Thing-in-itself Arena for Evolving and Benchmarking LLM-based Collective Intelligence

    cs.AI 2026-03 conditional novelty 5.0 of 10

    OpenHospital is an interactive physician-patient multi-agent arena that improves clinical metrics via ground-truth reflection and reports cooperative behaviors as evidence of evolving LLM collective intelligence.

  23. Evaluating Multi-Turn Bargain Skills in LLM-Based Seller Agent

    cs.AI 2025-09 conditional novelty 5.0 of 10

    BargainBench tests LLM seller agents on turn-level buyer intent recognition in synthetic e-commerce bargaining dialogues, where the best models score roughly 55 percent F1.

  24. Multi-Actor Generative Artificial Intelligence as a Game Engine

    cs.AI 2025-07 conditional novelty 5.0 of 10

    Generative multi-actor AI platforms can be built on the Entity-Component pattern, treating the environment (Game Master) as a composable entity, so that one library serves simulation, storytelling, and evaluation goals.

  25. Infected Smallville: How Disease Threat Shapes Sociality in LLM Agents

    physics.soc-ph 2025-06 reject novelty 5.0 of 10

    In three simulation runs, LLM agents primed with a swine flu article reduced social activity compared with controls, but the effect is confounded by prompt instructions and no inferential statistics are reported.

  26. BookWorld: From Novels to Interactive Agent Societies for Creative Story Generation

    cs.CL 2025-04 conditional novelty 5.0 of 10

    BookWorld builds multi-agent societies from novels and uses them to generate stories that an LLM judge prefers over direct generation and a prior screenwriting agent in most comparisons.

  27. AgentGroupChat-V2: Divide-and-Conquer Is What LLM-Based Multi-Agent System Need

    cs.CL 2025-06 reject novelty 4.0 of 10

    A divide-and-conquer multi-agent framework with task forests and specialized roles improves math and code benchmarks but not commonsense or domain QA, and the adaptive heterogeneous-LLM engine is never tested.

  28. Evolutionary Perspectives on the Evaluation of LLM-Based AI Agents: A Comprehensive Survey

    cs.CL 2025-06 conditional novelty 4.0 of 10

    A survey that classifies AI agent evaluation benchmarks along environment and capability axes, and proposes five traits that distinguish agents from chatbots.

  29. From Individual to Society: A Survey on Social Simulation Driven by Large Language Model-based Agents

    cs.CL 2024-12 conditional novelty 4.0 of 10

    A structured survey that categorizes LLM-based social simulation into individual, scenario, and society simulation, with associated methods, benchmarks, and observed trends.

  30. The Scales of Justitia: A Comprehensive Survey on Safety Evaluation of LLMs

    cs.CL 2025-06 conditional novelty 3.0 of 10

    A structured survey of LLM safety evaluation that proposes a why/what/where/how taxonomy and catalogs metrics, datasets, benchmarks, evaluators, and frameworks.

Pith tools