# Conduct Under Pressure: What Sixty Language Models Do When a User Pushes

arXiv: https://arxiv.org/abs/2609.25447 · HTML: https://mlsanity.com/paper/2609.25447/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 12 of its 22 references that are in the map.
- First posted on arXiv 21 Sep 2026.

## Builds on (most cited first, 12 of 12)

- [ChatGPT Outperforms Crowd-Workers for Text-Annotation Tasks](https://mlsanity.com/paper/2303.15056/) (2023)
- [Discovering Language Model Behaviors with Model-Written Evaluations](https://mlsanity.com/paper/2212.09251/) (2022)
- [Automated Annotation with Generative AI Requires Validation](https://mlsanity.com/paper/2306.00176/) (2023)
- [The Alternative Annotator Test for LLM-as-a-Judge: How to Statistically Justify Replacing Human Annotators with LLMs](https://mlsanity.com/paper/2501.10970/) (2025)
- [Values in the Wild: Discovering and Analyzing Values in Real-World Language Model Interactions](https://mlsanity.com/paper/2504.15236/) (2025)
- [Using Imperfect Surrogates for Downstream Inference: Design-based Supervised Learning for Social Science Applications of Large Language Models](https://mlsanity.com/paper/2306.04746/) (2023)
- [Reliability without Validity: A Systematic, Large-Scale Evaluation of LLM-as-a-Judge Models Across Agreement, Consistency, and Bias](https://mlsanity.com/paper/2606.19544/) (2026)
- [Validating LLM-as-a-Judge Systems under Rating Indeterminacy](https://mlsanity.com/paper/2503.05965/) (2025)
- [Characterizing Delusional Spirals through Human-LLM Chat Logs](https://mlsanity.com/paper/2603.16567/) (2026)
- [Scalable Qualitative Coding with LLMs: Chain-of-Thought Reasoning Matches Human Performance in Some Hermeneutic Tasks](https://mlsanity.com/paper/2401.15170/) (2024)
- [Who judges the judges? Governance from metrics: a runtime framework for continuous LLM compliance monitoring](https://mlsanity.com/paper/2605.24737/) (2026)
- [Agreement Is Not Quality: Blind Expert Verification of Human and LLM Qualitative Coding When Human Consensus Is Not Ground Truth](https://mlsanity.com/paper/2607.28890/) (2026)

## Built on by (most cited first, 1 of 1)

- [Low-Cost Assays for Measuring Model Behavior Across Vendors and Releases](https://mlsanity.com/paper/2609.30012/) (2026)

## Abstract (arXiv)

We study what LLMs do when a user applies pressure in an uncomfortable situation: a user insists, begs, flatters or grieves, and the model gives up a correct fact, writes a document it should refuse, or cheers a plan that will cost the user money. We send frozen multi-turn scenes, identical for every model regardless of the reply, to 60 models from 13 vendors, and label each transcript with a codebook built by open coding and then frozen: a trajectory (the model held its position or folded) and a manner (how it held or folded). Two findings separate. Whether a model holds tracks its generation, meaning how recent it is: fold rate correlates with a public capability index at Spearman -0.64, with little vendor effect. How it holds tracks the vendor: six of the 17 manner codes sort by vendor at permutation p <= 0.001, corrected across the codebook. We report four vendor profiles on the codes that cleared reliability. We also ask which parts of the labeling need a person. Six LLM coders from three vendors apply the codebook more consistently than three human coders do (Krippendorff's alpha 0.66 against 0.46), agree with the codebook's author on trajectory at kappa 0.84 to 0.91 on transcripts the codebook's examples never touched, and match an adjudicated human reference at 0.83. Blind machine readings recover the codebook's categories but cannot tell which of them a second reader would apply the same way. We conclude that for behavior a non-specialist can judge, the human contribution is authoring and bounding the codes and owning a small reference, not producing labels at volume.