x

LESSWRONG
LW

METR (org) — LessWrong

METR (org)

Edited by Ruby last updated 1st Jul 2024

Formerly ARC Evals

Posts tagged METR (org)

2

100METR's Observations of Reward Hacking in Recent Frontier Models

Daniel Kokotajlo

6mo

9

2

97Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity

5mo

43

2

10Review of METR’s public evaluation protocol

1y

0

1

242METR: Measuring AI Ability to Complete Long Tasks

Zach Stein-Perlman

8mo

106

1

153ARC Evals new report: Evaluating Language-Model Agents on Realistic Autonomous Tasks

2y

12

1

141METR's Evaluation of GPT-5

GradientDissenter

4mo

15

1

108Clarifying METR's Auditing Role

2y

1

1

90Introducing METR's Autonomy Evaluation Resources

Megan Kinniment, Beth Barnes

2y

0

1

70Interpreting the METR Time Horizons Post

8mo

13

1

67Reactions to METR task length paper are insane

8mo

43

1

65METR is hiring!

2y

1

1

64CoT May Be Highly Informative Despite “Unfaithfulness” [METR]

GradientDissenter

4mo

3

1

40ARC Evals: Responsible Scaling Policies

Zach Stein-Perlman

2y

10

1

30Improved visualizations of METR Time Horizons paper.

9mo

4

1

20How far along Metr's law can AI start automating or helping with alignment research?

Christopher King

9mo

21

Load More (15/19)

Add Posts