# Architecture-Sensitive Supervised Fine-Tuning for Screen-Conditioned Action Prediction: A PiSAR Benchmark

arXiv: https://arxiv.org/abs/2605.29400 · HTML: https://mlsanity.com/paper/2605.29400/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 14 of its 17 references that are in the map.
- First posted on arXiv 28 May 2026.

## Builds on (most cited first, 14 of 14)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Generative Agents: Interactive Simulacra of Human Behavior](https://mlsanity.com/paper/2304.03442/) (2023)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [Mind2Web: Towards a Generalist Agent for the Web](https://mlsanity.com/paper/2306.06070/) (2023)
- [CogAgent: A Visual Language Model for GUI Agents](https://mlsanity.com/paper/2312.08914/) (2023)
- [Using cognitive psychology to understand GPT-3](https://mlsanity.com/paper/2206.14576/) (2022)
- [UXAgent: An LLM Agent-Based Usability Testing Framework for Web Design](https://mlsanity.com/paper/2502.12561/) (2025)
- [Centaur: a foundation model of human cognition](https://mlsanity.com/paper/2410.20268/) (2024)
- [OPeRA: A Dataset of Observation, Persona, Rationale, and Action for Evaluating LLMs on Human Online Shopping Behavior Simulation](https://mlsanity.com/paper/2506.05606/) (2025)
- [Shop-R1: Rewarding LLMs to Simulate Human Behavior in Online Shopping via Reinforcement Learning](https://mlsanity.com/paper/2507.17842/) (2025)
- [Customer-R1: Personalized Simulation of Human Behaviors via RL-based LLM Agent in Online Shopping](https://mlsanity.com/paper/2510.07230/) (2025)
- [Not Yet AlphaFold for the Mind: Evaluating Centaur as a Synthetic Participant](https://mlsanity.com/paper/2508.07887/) (2025)
- [See, Think, Act: Online Shopper Behavior Simulation with VLM Agents](https://mlsanity.com/paper/2510.19245/) (2025)
- [Post-training makes large language models less human-like](https://mlsanity.com/paper/2605.07632/) (2026)

## Abstract (arXiv)

We benchmark three supervised fine-tuned models against frontier zero-shot baselines on a 661-row held-out slice of PiSAR (Persona, intent, Screen, Action, Rationale), a 12,929-tuple corpus of screen-anchored behavioural rationales curated from public app-store reviews, Pew American Trends Panel demographics, and the OPeRA shopper traces. Every model, frontier or fine-tuned, is evaluated on the same 661-row slice with the same scoring pipeline. Two findings. First, frontier zero-shot baselines (Claude Opus 4.7 and GPT-5.5) reach sem_sim 0.459 and 0.482 respectively; a fine-tuned Qwen3-VL-8B-Instruct reaches 0.783 and clears sem_sim >= 0.7 on 79% of rows, against 1-2% for either frontier baseline, a gap of 0.30 absolute on the same test set. Second, the same training data and recipe on Gemma-4-26B-A4B-IT scores only 0.441, in the same band as the frontier zero-shot baselines rather than the fine-tuned Qwen. We read this as a recipe-vs-model mismatch: the reasoning-tuned high-parameter model resists displacement and would likely need either more data or a stronger fine-tuning method.