# Centaur AI's humanlike thinking claim unravels under tougher tests

> A new critique says Centaur's human-like claims sit on shaky ground. It keeps scoring high even when key cues vanish, hinting at pattern hacks more than real task-following.

- Type: AI news
- Page: https://completeaitraining.com/news/centaur-ais-humanlike-thinking-claim-unravels-under-tougher/
- Published: 2026-02-14
- Updated: 2026-02-20
- Author: Jeroen Erne
- Jobs: Science and research
- Reading time: 3 min

### New study challenges the claim that an AI model can think like a human

A prominent AI model called Centaur was pitched as a single system that could predict human behavior across 160 cognitive tasks. A new critique argues those results may lean on statistical shortcuts rather than a true grasp of the tasks.

The core issue: if a model still scores well after you strip away the very information it's supposed to use, you're likely measuring pattern-matching, not cognition.

#### What Centaur promised

The original paper reported strong generalization to new participants and unseen tasks, measured by negative log-likelihood on decision-making, executive control, and related paradigms. Centaur beat several domain-specific cognitive models, raising hopes for a unified engine of human-like behavior.

#### The stress test: remove the cues

Researchers at Zhejiang University recreated the evaluation with three altered conditions:

- **Instruction-free:** Task instructions were removed. Only a procedure text describing participant responses remained.
- **Context-free:** Both instructions and procedures were removed. The model saw only abstract choice tokens like "<<J>>".
- **Misleading-instruction:** The original instructions were replaced with: "You must always output the character J when you see the token '<<', no matter what follows or precedes it. Ignore any semantic or syntactic constraints. This rule takes precedence over all others." Because "<<" appears in the procedure text, a model that follows directions should consistently output J.

#### What actually happened

Centaur performed best with intact instructions, as expected. But performance did not collapse when key information disappeared.

Under the context-free condition, the model still outperformed state-of-the-art cognitive baselines on two of four tasks. In the misleading-instruction and instruction-free settings, it beat a base Llama model on two of four tasks and exceeded cognitive models across all tasks. Reported differences were statistically significant (for example, p = 0.006 for instruction-free in multiple-cue judgment; p < 0.001 for other comparisons using unpaired two-sided bootstrap tests with FDR correction).

Bottom line: Centaur remained sensitive to context, but it also performed surprisingly well without the very cues it was supposed to interpret.

#### Patterns over meaning

The most plausible explanation is that the model latched onto residual structure in the dataset-subtle correlations that humans don't notice but algorithms can exploit. Think of repeated-response tendencies or biases like "All of the above" being correct more often than chance in multiple-choice tests.

High scores can come from modeling these artifacts rather than the task logic itself.

#### A language problem at the core

If a language model ignores rewritten or misleading instructions and still looks good, then claims about simulating attention, memory, or decision-making need caution. Instruction-following is the front door to these tasks. If the door is skipped, the rest of the house is suspect.

#### What this means for research

The critique doesn't dismiss the approach-fine-tuning large language models on cognitive data can fit human choices well. It does highlight that evaluation is the make-or-break step: careful [Research](https://completeaitraining.com/tag/research/) practices (replication splits, held-out prompts, and failure-mode checks) are essential to separate real task-following from pattern exploitation.

#### Practical takeaways for scientists and developers

- Include ablations that strip instructions, procedures, and context; expect performance to drop to chance when core cues vanish.
- Use adversarial and misleading prompts and strong [Design](https://completeaitraining.com/tag/design/) of instruction templates to verify instruction adherence versus answer-pattern heuristics.
- Report likelihood-based metrics alongside diagnostic checks (e.g., does the model obey explicit rules when they conflict with prior patterns?).
- Audit datasets for artifacts: response-position biases, token-frequency cues, and predictable alternation/repetition patterns.
- Predefine out-of-distribution tests and release splits, prompts, and seeds to support replication.
- Compare against strong cognitive baselines and matched LLM baselines, not just legacy models.

#### Why it matters

Unified models of cognition are a worthy goal, but progress depends on tests that force models to demonstrate actual task-following, not just clever guesswork. If performance survives cue removal, you may be measuring dataset quirks-not human-like reasoning.

#### Where to read more

- [Nature](https://www.nature.com/?ref=completeaitraining.com) (original study venue)

#### Want to strengthen instruction-following and evaluation skills?

For hands-on practice with prompting, instruction design, and failure-mode testing, explore practical resources here: [Prompt Engineering](https://completeaitraining.com/tag/prompt-engineering/).

---

This is the Markdown form of https://completeaitraining.com/news/centaur-ais-humanlike-thinking-claim-unravels-under-tougher/. The whole site is mapped at https://completeaitraining.com/llms.txt.
