GRAFT achieves 0.3866 co-bps SOTA on MC Maze and supports cross-day recalibration by updating 9.21% of parameters while retaining competitive performance on scaled datasets.
and Pandarinath, Chethan , title =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
GRAFT: Gain-Recalibrated Adapters for Transformer-Based Neural Population Activity Modeling
GRAFT achieves 0.3866 co-bps SOTA on MC Maze and supports cross-day recalibration by updating 9.21% of parameters while retaining competitive performance on scaled datasets.