AI brief
A SPAR group ran a 29 × 29 propensity SFT spillover experiment testing how fine-tuning for one propensity changes each of 29 other propensities.
Written by AI from LessWrong's published text.
Source
“As in you specifically. tldr: My SPAR group ran a 29 × 29 propensity SFT spillover experiment (i.e. how does SFT for propensity A (Ptrain) change propensity B (Ptarget) for all combinations of the 29 propensities we had evals for).”
Interpretation or community commentary rather than straight reporting.
Read original story ↗