# Quantifying the Utility of User Simulators for Building Collaborative LLM Assistants

arXiv: https://arxiv.org/abs/2605.09808 · HTML: https://mlsanity.com/paper/2605.09808/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Builds on 88 of its 105 references that are in the map.
- First posted on arXiv 10 May 2026.

## Builds on (most cited first, 30 of 88)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Deep reinforcement learning from human preferences](https://mlsanity.com/paper/1706.03741/) (2017)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [Generative Agents: Interactive Simulacra of Human Behavior](https://mlsanity.com/paper/2304.03442/) (2023)
- [Qwen Technical Report](https://mlsanity.com/paper/2309.16609/) (2023)
- [Constitutional AI: Harmlessness from AI Feedback](https://mlsanity.com/paper/2212.08073/) (2022)
- [UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction](https://mlsanity.com/paper/1802.03426/) (2018)
- [G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment](https://mlsanity.com/paper/2303.16634/) (2023)
- [Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems](https://mlsanity.com/paper/2005.01643/) (2020)
- [Generative Adversarial Imitation Learning](https://mlsanity.com/paper/1606.03476/) (2016)
- [Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments](https://mlsanity.com/paper/1706.02275/) (2017)
- [Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real World](https://mlsanity.com/paper/1703.06907/) (2017)
- [A Survey on LLM-as-a-Judge](https://mlsanity.com/paper/2411.15594/) (2024)
- [τ-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains](https://mlsanity.com/paper/2406.12045/) (2024)
- [Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference](https://mlsanity.com/paper/2403.04132/) (2024)
- [Scaling Laws for Reward Model Overoptimization](https://mlsanity.com/paper/2210.10760/) (2022)
- [Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Models](https://mlsanity.com/paper/1805.12114/) (2018)
- [When to Trust Your Model: Model-Based Policy Optimization](https://mlsanity.com/paper/1906.08253/) (2019)
- [Sim-to-Real Transfer of Robotic Control with Dynamics Randomization](https://mlsanity.com/paper/1710.06537/) (2017)
- [WildChat: 1M ChatGPT Interaction Logs in the Wild](https://mlsanity.com/paper/2405.01470/) (2024)
- [Robust Adversarial Reinforcement Learning](https://mlsanity.com/paper/1703.02702/) (2017)
- [Prometheus 2: An Open Source Language Model Specialized in Evaluating Other Language Models](https://mlsanity.com/paper/2405.01535/) (2024)
- [LLMs Get Lost In Multi-Turn Conversation](https://mlsanity.com/paper/2505.06120/) (2025)
- [Scaling Synthetic Data Creation with 1,000,000,000 Personas](https://mlsanity.com/paper/2406.20094/) (2024)
- [Rubrics as Rewards: Reinforcement Learning Beyond Verifiable Domains](https://mlsanity.com/paper/2507.17746/) (2025)
- [Quantifying Generalization in Reinforcement Learning](https://mlsanity.com/paper/1812.02341/) (2018)

## Built on by (most cited first, 10 of 10)

- [SWE-Together: Evaluating Coding Agents in Interactive User Sessions](https://mlsanity.com/paper/2606.29957/) (2026)
- [Learning User Simulators with Turing Rewards](https://mlsanity.com/paper/2606.19336/) (2026)
- [One Frozen Simulator Is Not Enough: Simulator Collapse in Multi-Agent RL](https://mlsanity.com/paper/2608.12253/) (2026)
- [Beyond expert users: agents should help users construct preferences, not just elicit them](https://mlsanity.com/paper/2606.30863/) (2026)
- [Measuring the Behavioral Fidelity of Long-Horizon Human Activity Simulations](https://mlsanity.com/paper/2609.01257/) (2026)
- [Proactive Dialogue Policy Optimization via Cognitive-State Transition](https://mlsanity.com/paper/2609.34948/) (2026)
- [CUEing User Simulators: Calibrated User Embeddings for Multi-Turn Benchmarking](https://mlsanity.com/paper/2610.02460/) (2026)
- [Dataset Signatures in Human-LLM Interactions and User Modeling](https://mlsanity.com/paper/2610.05534/) (2026)
- [MIMESIS: Learning User Simulators as Training Environments for Interactive Agents](https://mlsanity.com/paper/2610.09484/) (2026)
- [Training User Simulators with Latent Intrinsic Feedback](https://mlsanity.com/paper/2610.13617/) (2026)

## Abstract (arXiv)

User simulators are increasingly leveraged to build interactive AI assistants, yet how to measure the quality of these simulators remains an open question. In this work, we show how simulator quality can be quantified in terms of its downstream utility: how an LLM assistant trained with this user simulator performs in the wild when interacting with real humans. In a controlled experiment where only the user simulator varies, we train LLM assistants via reinforcement learning against a spectrum of simulators, from an LLM prompted to role-play a user to one fine-tuned on human utterances from WildChat. As evaluation, we measure pairwise win rates in a user study with 283 participants and on WildBench, a benchmark derived from real human--AI conversations. Training against the role-playing LLM yields an assistant statistically indistinguishable from the initial assistant in our user study (51% win rate), whereas training against the fine-tuned simulator yields significant gains (58% over the initial and 57% over the one trained against role-playing). Closer inspection reveals three further patterns: methods for making role-playing LLMs more realistic (e.g., persona conditioning) improve trained assistants but do not close the gap to the fine-tuned simulator; scaling the simulator's model size benefits the fine-tuned simulator but yields no gain for role-playing ones; and assistants trained against role-playing simulators fail to generalize when paired with other simulators at test time, while the one trained against fine-tuned simulator does. Together, these results argue for grounding user simulators in real human behavior and measuring their quality by their downstream effect on real users.