Individual results

View docs

View 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 subject relative clause)
Item
Condition
intro np_subject that embed_vp the embed_np matrix_v reflexive
Item Condition intro np_subject that embed_vp the embed_np matrix_v reflexive
1 match_sing The author that liked the senators hurt herself
1 mismatch_sing The author that liked the senators hurt themselves
1 match_plural The authors that liked the senator hurt themselves
1 mismatch_plural The authors that liked the senator hurt herself
Prediction performance for Vanilla LSTM on Reflexive Number Agreement (feminine; with subject relative clause)
Accuracy
Formula
Description
AccuracyPredictionDescription
0.00% (664,match_sing/8,reflexive) < (666,mismatch_sing/8,reflexive) No description provided.
100.00% (665,match_plural/8,reflexive) < (663,mismatch_plural/8,reflexive) No description provided.