SOURCE-LINKED INTELLIGENCE

Replicating TRACE: A Practitioner's Guide to Its Threshold and Particle Budget

arXiv · AI, language, vision and robotics · article · Sep 1, 2026 · UTC

TRACE (Math & Lienhart, arXiv:2602.01135) reads causal graphs over event types out of a pretrained autoregressive sequence model by thresholding a per-position conditional-mutual-information estimate at a fixed tau. We independently replicate its headline synthetic result: with tau selected on a validation split, mean per-sequence F1 against exact interventional truth reaches 0.90-0.91 at vocabulary size 1000 (paper: 0.91) and 0.86-0.91 from 100 to 2000. First, the optimal threshold is pinned to the truth margin, not to any constant: at every size the errors at tau* straddle the delta = 0.05 m

Read original source ↗ Open in workspace

recordType

paper

region

Global