Induction Heads in Transformers Quiz

Answers and explanations appear as you take the quiz.

1. What rule does an induction head essentially implement?
   • Predict the most frequent token in the language
   • Always copy the previous token unchanged
   • Randomly shuffle the context tokens
   • See [A][B] earlier, then on [A] again predict [B]
2. What capability was linked to induction-head behavior in the cited 2022 study?
   • In-context learning in the examined models, with broader large-model claims framed as hypotheses
   • A proof that every form of reasoning uses only one identical circuit
   • Changing the tokenizer vocabulary during every prompt
   • Permanent weight updates whenever a repeated token appears
3. What role does the 'previous-token head' play in the induction circuit?
   • It generates the final softmax
   • It removes positional information
   • It copies each token's predecessor information forward so matching can happen
   • It normalizes attention weights
4. How many attention heads are typically involved in the canonical induction circuit?
   • Exactly one
   • Two, composed across layers
   • Hundreds in a single layer
   • Zero; it is feed-forward only
5. What training relationship did the induction-head study report in the models it examined?
   • Head formation coincided with changes in loss behavior and increased in-context learning
   • All heads were fully formed before any training
   • Induction behavior could only appear after every parameter was frozen forever
   • The same exact training transition must occur in every future model
6. Which task would an induction head handle most directly?
   • Computing a numerical gradient
   • Compressing the model's weights
   • Encrypting the input prompt
   • Repeating a sequence by copying what followed earlier matching tokens
7. What roles do query-key matching and the output-value path play in the canonical induction circuit?
   • Query-key computations generate the final text filename; output-value computations train the tokenizer
   • Query-key computations select the matching context position; output-value computations promote the token information there
   • Both paths only count the total number of context words
   • The output-value path selects a position without any attention pattern
8. Does a demonstrated two-head induction circuit prove that every large-model in-context task uses that mechanism alone?
   • Yes; every architecture and prompt has already been causally tested
   • Yes; no other network computation can influence a prediction
   • No; the wider explanation remains a hypothesis needing model- and task-specific evidence
   • Only the number of parameters determines whether the proof is universal