OpenInterpretability
Collection
Agent-safety interpretability arc, first public Qwen3.6 SAEs, the probe/guard suite, and the agent-trajectory benchmarks behind them. openinterp.org • 37 items • Updated
The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
Labeled reasoning rollouts used to train/evaluate the ReasoningGuard linear probe (reasoning-integrity monitoring) on Qwen3.6-27B, part of the OpenInterpretability probe stack.
Companion to the probe weights: caiovicentino1/ReasoningGuard-linearprobe-qwen36-27b.
Part of the OpenInterpretability conditionally-causal probe program.
@techreport{vicentino2026probes,
title = {Conditionally-Causal Probes: Five Operational Constraints on Linear-Probe Causality in Qwen3.6-27B},
author = {Vicentino, Caio}, year = {2026}, institution = {OpenInterpretability},
url = {https://openinterp.org/research/papers/conditionally-causal-probes}
}
Full research index: https://openinterp.org/research