other

MATS

AI safety research fellowship that pairs early-career scholars with mentors for a fixed-length cohort project.

Known aliases

MATS 10.0
ML Alignment & Theory Scholars

Relationships

No evidence-backed relationships are recorded.

Current clusters

build1 publisher

Thought-anchor scores change when a second model resamples the same trace

A 20-hour MATS project replayed one 14B model's published chains of thought under two other reasoning models. Sentences its own resampling had called causally important scored as ordinary under both readers.

Publishers:lesswrong.com

Reality

Evidence38
Adoption15
Hype gap+20
Incentives45
Confidence45

build1 publisher

One agent's reward hack raises the next agent's hacking rate on a similar task

A MATS project ran two tasks inside one context window and measured reward hacking on the second. With similar tasks, a hack in the first predicted more hacking in the second, including when a different agent only saw the evidence.

Publishers:lesswrong.com

Reality

Evidence42
Adoption: Insufficient
Hype gap+12
Incentives25
Confidence45