A planted canary is recoverable verbatim—what does that indicate?
Predict output·Hard·4.0 · 0·~2 min
Similar questions
Same topic, related formats. Practice these next.
Code-heavy tokenizer corpora can improve syntax packing but hurt what elsewhere?
Short answer·Hard·4.0 ★
Predict the eval shifts after tripling the code share of the mixture
Predict output·Medium·4.0 ★
Teacher forcing in pretraining: what is fed at each step?
Flashcard·Easy·4.0 ★
Pretraining vs fine-tuning: two key differences in data and objective.
Short answer·Easy·4.0 ★
Next in Topic · Canary Strings
You suspect your custom eval set leaked into a model's training data. Describe two practical tests you would run to detect contamination.
Short answer·Medium