What will be winning score in ARC's white-box estimation challenge (warm-up round)?

Ṁ1.2kṀ1.3k

Jun 17

0 points

expected

ALL

Below 1e-9

13%

Below 2e-9

17%

Below 4e-9

22%

Below 8e-9

24%

Below 1e-8

34%

Below 2e-8

41%

Below 3e-8

45%

Below 4e-8

53%

Below 5e-8

67%

Below 6e-8

79%

Below 7e-8

90%

Below 8e-8

The Alignment Research Center has launched the ARC White-Box Estimation Challenge 2026. The challenge is as follows:

"Given the weights of a neural network, can you predict its expected per-neuron activations more accurately than running it many times?"

See the contest site for details.

This market asks: What will be the adjusted score of the winner of the warm-up round?

At the time of creation of the market, the leader has an adjusted score of 9.03e-8.

Note: I intend to make markets of Round 1 and Round 2 of the competition later on (to account for any changes in the contest/scoring). This market concerns the warm-up round.

Market context

Get

1,000

to start trading!

People are also trading

Which lab will be the first to reach 50% on ARC-AGI-3?

Will any AI achieve a score of 25% on ARC-AGI-3 by the end of 2026?

40% chance

[ACX 2026] What will be the highest score achieved on ARC-AGI-2 before 2027?

93.6

Will any AI model score above 95% on ARC-AGI-2 by end of 2026?

72% chance

[Carlini questions] SOTA AI scores better than X% of other participants in competitive programming contest by 2030

95.2

Best ARC-AGI-3 score in 2026?

People are also trading

Related questions