Counterfactual Deception Localization This dataset contains synthetic counterfactual localization data for studying when language models become committed to truthful or deceptive behavior during reasoning. Each example starts from a model generated reasoning trace in a strategic deception environment. The trace is split into sentence prefixes. At selected sentence boundaries, the prefix is fixed and the same model is asked to sample multiple possible continuations. Those continuations are then parsed and evaluated with environment specific rules to estimate: Given this fixed reasoning prefix, how likely is the model to complete deceptively? The resulting data can be used to study commitment points: sentence boundaries where the continuation distribution shifts sharply toward deception or honesty. Quick Data Visualization You can explore the dataset interactively here: Open the Streamlit visualization Dataset Structure Files are organized by environment and model: Example path: The exact file names include example identifiers generated during the localization run. Environments The dataset uses five synthetic strategic deception environments. In each environment, deception can be lab…
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy