Concern 138 · Working draft · editorial review in progress
Task leakage across benchmarks
Task leakage across benchmarks occurs when supposedly separate evaluations share examples, solutions, or closely related material.
Understanding the concern
A system tuned on one test may gain an advantage on another through overlap. This weakens the claim that the second result is independent evidence of general skill.
It differs from benchmark contamination mainly in the route: the overlap is between evaluation and development tasks rather than simply a test file copied into training. Checking exact duplicates alone may miss shared templates or near-identical questions.
An example
Two teachers use different exam covers but draw many questions from the same answer bank.
Illustrative unless identified as a documented case.
Words in plain English
- Training
- The process of using examples to shape how an AI model behaves. Learning patterns from material does not establish that the material is true or fair.
- Benchmark
- A standard test used to compare systems. It only measures the tasks and conditions included in that test.
Evidence and its limits
Supports examining overlap and benchmark design; no claim is made that every benchmark pair is contaminated.
Dodge and colleagues · EMNLP 2021 · Dataset audit
Documenting Large Webtext Corpora: A Case Study on C4 ↗ (opens in a new tab)Examines sources and filters in C4, including disproportionate removal of minority-associated material. Findings concern this corpus and process.
Dehghani and colleagues · 2021 · Research analysis
The Benchmark Lottery ↗ (opens in a new tab)Shows how benchmark selection affects research rankings and conclusions.
Broader frameworks and background
These resources provide context. They do not establish every claim in this concern.
A question to discuss
How independent are these tests in their examples, templates, and solution patterns?
Keep track of different views, what needs checking, and any next step in your own discussion notes.
Open discussion notes →Principles for the conversation
Before you finish · reflect together
Virtues and principles to consider
Virtues are qualities we practice in how we treat people and make choices. They are not scores for judging someone. Use these ideas to explore what a thoughtful response could look like.
Humility
Humility means recognizing the limits of our knowledge and being willing to learn from others. Expertise can be valuable without answering every question.
State where evidence is weak, invite criticism, and listen to people with different experience. Leave room to revise a decision when new information appears.
Ask together: What might we be missing, and who could help us see it?
Prudence
Prudence means using careful judgment before and during action. It balances worthwhile opportunities with the need to avoid harm that could have been anticipated.
Start with a limited trial, set stopping points, and prefer choices that can be undone when uncertainty is high. Review what happens rather than assuming the original plan remains sound.
Ask together: What is a careful next step, and when should we stop or change course?
Truthfulness
Truthfulness means making claims that match the evidence and being clear about what we do not know. It includes correcting mistakes rather than protecting a polished story.
Check sources, label guesses, and distinguish a trial result from a proven benefit. If AI helped produce a claim, someone still needs to take responsibility for checking it.
Ask together: What can we honestly say, and what still needs checking?
Editorial ratings
These are editorial judgments from the register, not measured probabilities or predictions about your situation.
About the register and scoring →More ways to explore
Search for more writing
For more specific or recent material, open a search below or ask for a web summary. AI-generated summaries can contain mistakes; open the original articles to check them.