Individual results
View docsView in-depth performance of a single language model on a single test suite.
Region-by-region surprisal
Sample item for Reflexive Number Agreement (feminine; with prepositional phrase)
The first item of the test suite is shown below for quick reference. Please visit the page for Reflexive Number Agreement (feminine; with prepositional phrase) to see the full list of items.
Item |
Condition
|
intro | np_subject | prep | the | prep_np | matrix_v | reflexive |
---|---|---|---|---|---|---|---|---|
Item | Condition | intro | np_subject | prep | the | prep_np | matrix_v | reflexive |
1 | match_sing | The | author | next to | the | senators | hurt | herself |
1 | mismatch_sing | The | author | next to | the | senators | hurt | themselves |
1 | match_plural | The | authors | next to | the | senator | hurt | themselves |
1 | mismatch_plural | The | authors | next to | the | senator | hurt | herself |
Prediction performance for GPT-2 XL on Reflexive Number Agreement (feminine; with prepositional phrase)
Accuracy |
Formula
|
Description |
---|---|---|
Accuracy | Prediction | Description |
57.89% | (564,match_sing/7,reflexive) < (566,mismatch_sing/7,reflexive) | No description provided. |
94.74% | (565,match_plural/7,reflexive) < (563,mismatch_plural/7,reflexive) | No description provided. |