# PsyMemBehav: Psychology-Informed and Memory-Grounded Two-Level Identity-to-Action Reasoning for Social-Media Behavior Simulation

arXiv: https://arxiv.org/abs/2610.14570 · HTML: https://mlsanity.com/paper/2610.14570/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Reasoning benchmarks and evaluation, one of 17 directions in LLM reasoning.
- Builds on 19 of its 60 references that are in the map.
- First posted on arXiv 10 Oct 2026.

## Placed in

- LLM reasoning / [Reasoning benchmarks and evaluation](https://mlsanity.com/lr/reasoning-benchmarks-and-evaluation/)

## Builds on (most cited first, 19 of 19)

- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities](https://mlsanity.com/paper/2507.06261/) (2025)
- [Constitutional AI: Harmlessness from AI Feedback](https://mlsanity.com/paper/2212.08073/) (2022)
- [G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment](https://mlsanity.com/paper/2303.16634/) (2023)
- [DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence](https://mlsanity.com/paper/2606.19348/) (2026)
- [RLAIF vs. RLHF: Scaling Reinforcement Learning from Human Feedback with AI Feedback](https://mlsanity.com/paper/2309.00267/) (2023)
- [Prometheus: Inducing Fine-grained Evaluation Capability in Language Models](https://mlsanity.com/paper/2310.08491/) (2023)
- [Character-LLM: A Trainable Agent for Role-Playing](https://mlsanity.com/paper/2310.10158/) (2023)
- [Large Language Models Empowered Agent-based Modeling and Simulation: A Survey and Perspectives](https://mlsanity.com/paper/2312.11970/) (2023)
- [Ex Machina: Personal Attacks Seen at Scale](https://mlsanity.com/paper/1610.08914/) (2016)
- [CharacterEval: A Chinese Benchmark for Role-Playing Conversational Agent Evaluation](https://mlsanity.com/paper/2401.01275/) (2024)
- [MOSAIC: Modeling Social AI for Content Dissemination and Regulation in Multi-Agent Simulations](https://mlsanity.com/paper/2504.07830/) (2025)
- [BASES: Large-scale Web Search User Simulation with Large Language Model based Agents](https://mlsanity.com/paper/2402.17505/) (2024)
- [GA-S3: Comprehensive Social Network Simulation with Group Agents](https://mlsanity.com/paper/2506.03532/) (2025)
- [Can LLMs Simulate Social Media Engagement? A Study on Action-Guided Response Generation](https://mlsanity.com/paper/2502.12073/) (2025)
- [AdaMARP: An Adaptive Multi-Agent Interaction Framework for General Immersive Role-Playing](https://mlsanity.com/paper/2601.11007/) (2026)
- [PsyMem: Fine-grained psychological alignment and Explicit Memory Control for Advanced Role-Playing LLMs](https://mlsanity.com/paper/2505.12814/) (2025)
- [RippleMem: From Isolated Retrieval to Associative Recollection for Long-Term Agent Memory](https://mlsanity.com/paper/2608.13334/) (2026)
- [DREAM: LLM-based Dynamic Role-playing via Event-Aware Memory Graph](https://mlsanity.com/paper/2608.05170/) (2026)

## Abstract (arXiv)

Large language models (LLMs) support social simulation, but faithfully reproducing real users requires grounding both structured platform decisions and language in sparse, noisy behavioral evidence. We introduce PsyMemBehav, which maps psychology-informed profiles and retrieved memories to target-first platform behavior through two-level identity-to-action reasoning. To support this approach with comprehensive behavioral traces, we construct a dataset PsyMemBehav-bench from 5.57M observed behaviors of 5,862 real users (approximately 950 per user) across Reddit, Bluesky, and Twitter. Each recorded engagement is modeled as target selection followed by Comment, Like, or Repost and optional response generation. PsyMemBehav combines 41-dimensional user descriptors, top-5 BM25 memories, reverse-CoT supervision, and rubric-guided DPO over complete reasoning-action trajectories. Experiments show that PsyMemBehav 27B ranks first in 11 of 12 reported individual-level metrics, improving average response quality over GPT-5.4 from 54.28 to 59.68 (9.95% relative), and obtains the lowest group-level MAE (0.081) and MSE (0.009) among simulated systems. These results support evidence-grounded user modeling for observed-thread continuation; the benchmark remains conditional on recorded engagement and does not model no-action decisions.