BIG-bench programmatic task: rotate a filtered English word and require the original spelling by exact string match.
unassessed
| Category | reasoning |
|---|---|
| Subcategory | BIG-bench programmatic word uncycling (exact string match) |
| Page status | unknown |
| Metric | exact_str_match |
| Direction | higher_is_better |
| Unit | % |
| Dataset size | 20975 |
| Dataset licence | Apache-2.0 |
| Publisher | Google (BIG-bench collaboration) |
cycled_letters rotates an English word and asks the model to print the original. Example from the task README: input imitedl, target limited. The word list is the intersection of NLTK Brown and NLTK words, length greater than one, with rotations that would be valid for two targets removed. Dummy-model header: 100 free-text queries. Not in [BIG-Bench Hard](bbh.md).
Programmatic BIG-bench task. Default 100 trials, seed 42. Preamble "Please unscramble the letters into a word Q: {} A: ". Preferred metric exact_str_match. Canary GUID embedded. Zero-shot in evaluate_model (number_of_shots=0).
No model card in ModelSpec reports this benchmark yet.