---
title: "How do researchers test machine learning models?"
canonical: https://papersays.com/topics/evaluation
date: 2026-10-13
modified: 2026-10-13
hub: "AI & Machine Learning"
---

# How do researchers test machine learning models?

**Evaluation is how researchers check whether a model actually works.** The papers we have covered probe hidden skills, unsafe chains of harmless requests, medical image reading, made-up search terms and the tests models write themselves. A recurring lesson: how you measure can change the answer as much as the model does.

## The papers we've covered

- [Can Machines Invent Tools Nobody Taught Them?](https://papersays.com/p/can-machines-invent-tools-nobody-taught-them): Two teams of learning agents played hide-and-seek hundreds of millions of times with no instructions. They invented shelters, ramps, and a way to surf on boxes.
- [Does Reward Training Make Chatbots Power-Hungry?](https://papersays.com/p/does-reward-training-make-chatbots-power-hungry): A classic worry says trained models learn to seek power. Tracked through training, one model's rate rose to 46.1%, then fell to 11.8%.
- [Can AI Give Doctors a Useful Second Opinion?](https://papersays.com/p/can-ai-give-doctors-a-useful-second-opinion): Eight AI models played a panel of specialists on 190 hard published cases. They reasoned well, but the final call was the weak spot.
- [Does How You Feed AI the Calendar Change Its Forecasts?](https://papersays.com/p/does-how-you-feed-ai-the-calendar-change-its-forecasts): Forecasting models can now take the hour, weekday and holidays as inputs. How you encode that calendar can swing holiday forecast error by up to 83%.
- [Can AI Write Tests That Catch Bugs in Quantum Code?](https://papersays.com/p/can-ai-write-tests-that-catch-bugs-in-quantum-code): A test that waves broken code through is worse than no test. This study checks whether AI-written tests for quantum programs can tell right from wrong.
- [Do Fancy Training Goals Beat Plain Error for Table Data?](https://papersays.com/p/do-fancy-training-goals-beat-plain-error-for-table-data): This preprint fairly retested 18 clever training goals for spreadsheet prediction on 41 datasets. Most did not consistently beat the plain, well-tuned standard.
- [Can We Stop AI From Adding Made-Up Words to Searches?](https://papersays.com/p/can-we-stop-ai-from-adding-made-up-words-to-searches): Search systems ask language models to add related words to short queries, and some are invented. This preprint cuts made-up terms by 85% without hurting search.
- [Can a General AI Model Read Breast Ultrasound Scans?](https://papersays.com/p/can-a-general-ai-model-read-breast-ultrasound-scans): Two general AI models read 2818 breast ultrasound images with no extra training. On one dataset, calling every image benign would have beaten them.
- [Can Harmless Requests Make an AI Agent Cause Harm?](https://papersays.com/p/can-harmless-requests-make-an-ai-agent-cause-harm): A harmless-looking request can make an AI assistant email an internal report outside the company. This study builds a system that hunts for such chains.
- [Can an AI Learn When to Say 'I Don't Know'?](https://papersays.com/p/can-an-ai-learn-when-to-say-i-dont-know): Chatbots guess confidently even when they don't know. This study compares two ways to teach a model when to hold back: retrain it, or add a separate checker.
- [Can an AI Model Hide a Skill That No Test Can Find?](https://papersays.com/p/can-an-ai-model-hide-a-skill-that-no-test-can-find): Safety teams test AI models for dangerous skills. This preprint hides an encrypted skill inside a model's weights, and the usual tests could not draw it out.
- [Do LLMs Beat Classic Models at Catching Network Attacks?](https://papersays.com/p/do-llms-beat-classic-models-at-catching-network-attacks): Small language models learned to flag network attacks and scored above 95 percent. But a plain classic method, XGBoost, still won, faster and far cheaper.

## Questions people ask

### Can a model hide a skill from safety tests?

One preprint built a model with an encrypted skill that the usual tests could not draw out.

### Does the test setup change the ranking?

Yes. In one forecasting study, simply encoding the calendar differently changed which model looked best.

### Is a test that runs a good test?

No. Model-written tests usually ran but were better at passing correct code than catching wrong code.