Winograd

PaLM 2-L 89.5% Dropped

Models pick which noun an ambiguous pronoun refers to in Winograd schema sentence pairs that need world knowledge. Scored by accuracy.

Models passed 90 percent by 2019, which led to the larger WinoGrande dataset.

Made by Hector Levesque , Ernest Davis , and Leora Morgenstern Commonsense All tasks public Last in a launch post May 2023 , dropped after it saturated Last model added May 2023

Level: Everyday Most adults, without training
Scored: Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
Answers: Multiple choice

cs.nyu.edu

Model scored in lab cards

Launch post lists it, from 1 lab

PaLM 2-L leads with 89.5%

Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.

Bars Table

89.5 PaLM 2-L

1 PaLM 2-L 89.5%

#	Model	Score	Setting	Date	Source
1	PaLM 2-L	89.5%		May 2023	Model card

The best score passed 85% in May 2023

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Reported by the lab in its model card

Labs dropped it from launch posts once the best score passed 85%

Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.

Google DeepMind

Date	Launch post	Score	As the post writes it
10 May 2023	PaLM 2-L	89.5	Winograd, 1-shot

PaLM 2-L leads with 89.5%

The best score passed 85% in May 2023

Labs dropped it from launch posts once the best score passed 85%

More in Knowledge