Gids-gekoppelde quiz · Middelmatig Niveau

Benchmarkverontreinigingsquiz

Test hoe benchmarkitems in trainingsgegevens lekken, detectiemethoden zoals n-gram overlap en Min-K% Prob, en oplossingen zoals canaries en tijdsverschillen.

Gerelateerde gidspadenBenchmarkverontreiniging
Vraag 1 van 8

Wat is benchmarkverontreiniging?

Bekijk elke vraag8 vragen

Antwoorden en uitleg verschijnen terwijl je de quiz doet.

  1. Wat is benchmarkverontreiniging?
    • When a benchmark contains incorrect answer keys
    • When a model is tested on too few questions
    • When benchmark test items or their answers appear in a model's training data
    • When two benchmarks measure the same skill
  2. Wat is het doel van de kanarie-snaar in BIG-bench-taken?
    • To let data teams filter those documents from training data and test whether a model saw them
    • To encrypt the answers
    • To mark which tasks are the hardest
    • To identify which model generated each answer
  3. Wat suggereerde het GSM1k-onderzoek?
    • All models scored higher on the new problems
    • GSM8K has no errors
    • Contamination cannot be measured without training data
    • Some models scored noticeably lower on fresh, matched problems than on GSM8K
  4. Welke overlapanalyse is uitgevoerd in het GPT-3-paper?
    • Embedding similarity on full documents
    • 13-gram overlap between benchmarks and training data
    • Human review of every training document
    • Min-K% Prob on each benchmark
  5. Wat is het kernidee van Min-K% Prob?
    • Removing the least common k percent of words from training data
    • Training a second model to classify contaminated text
    • Text the model has seen tends to have few very low-probability tokens
    • Measuring the length of the model's answers
  6. How does a time-based split, as in LiveCodeBench, help detect contamination?
    • It compares performance on problems published before and after the model's training cutoff
    • It measures how long the model takes to answer
    • It splits the model into older and newer layers
    • It only counts problems answered within a time limit
  7. Why does finding no exact overlap not prove a benchmark is clean?
    • Exact matching is too slow to run
    • Benchmarks change their answers over time
    • Models never memorize exact text
    • Paraphrased or translated copies can slip past string matching
  8. In n-gram decontamination, what happens if n is chosen too small?
    • Near-copies with small edits are missed
    • Many false positives from common phrases are flagged
    • The filter stops working entirely
    • Only answers, not questions, are matched