test_set sha256 eb175779d1d5adf99c893ff1f82de9b18ada912d3f385b3f5419195afdfb374a bytes 2566199 entries 200 id 100 ood 100 example 3 top-level fields [('scenario', 200), ('topic', 200), ('key_characters', 200), ('dialogues', 200), ('speaking_characters_w_env', 200), ('major_characters', 200), ('weight', 200), ('plot', 200), ('character_profiles', 200), ('book', 200), ('i_p', 200), ('i_c', 200), ('tag', 200)] unique keys 200 tag values Counter({'id': 100, 'ood': 100}) id ∪ ood == test_set True id∩ood 0 test_set order == id + ood True first100 are id-file True content equal to id/ood entries 0 books total 101 books id 69 books ood 32 book overlap 0 key_characters fields [('name', 456), ('thought', 456)] example file key_characters fields [('name', 8), ('motivation', 8)] id file key_characters fields [('name', 238), ('thought', 238)] key characters with non-empty thought 456 of 456 conversations where a speaking character has no key_characters entry 25 speaking lists containing Environment 130 speaking[0]==dialogues[0].character 154 speaking[0]==Environment 4 dialogues length min/mean/max 9 12.89 32 dialogue longer than 20 lines 7 reference messages with [..] 2281 of 2578 major characters mean 2.32 major_characters containing Environment 0 speaking mean 3.045 CJK characters in dialogues 0 profiles empty for a speaking character 55 plot.key_characters description field present Counter({True: 655}) plot summary empty 0 i_p fields equal 200