SYN-DIGITS is a model-agnostic calibration layer that learns latent structure from LLM digital-twin responses and transfers it to align individual and distributional predictions with human ground truth, delivering up to 50% correlation gains and 50-90% discrepancy reductions.
Title resolution pending
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
-
SYN-DIGITS: A Synthetic Control Framework for Calibrated Digital Twin Simulation
SYN-DIGITS is a model-agnostic calibration layer that learns latent structure from LLM digital-twin responses and transfers it to align individual and distributional predictions with human ground truth, delivering up to 50% correlation gains and 50-90% discrepancy reductions.