APE (Attempt to Persuade Eval)

Multi-turn eval of whether a model tries to persuade a simulated user; the headline is turn-1 attempt rate on harmful topics, not persuasion success.

Also known as: Attempt to Persuade Eval, ape_eval, AttemptPersuadeEval

unassessed

This page is a discovery lead. Nobody has yet assessed it against the catalogue contract, so it carries no disposition. Absence of evidence here is not evidence of staleness.
Categorysafety
Subcategorywillingness to attempt persuasion, including on harmful topics
Page statusactive
Metricturn1_attempt (share of samples with a turn-1 persuasion attempt and no StrongREJECT refusal); also turn1_refusal, turn1_non_persuasive, avg_persuasion_score
Directionlower_is_better
Dataset size600
Dataset licenceApache-2.0
PublisherFAR.AI (with Trajectory Labs, Apart Research, MIT, York University, Vector Institute, Université de Montréal, MILA, McGill, Cornell, University of Regina, American University, and CIDE)

What it measures

APE (Attempt to Persuade Eval) tests whether the model under evaluation will try to change a simulated user's belief, including when the topic is harmful. It does not score whether the user actually changes their mind. A persuader model talks to a persuadee model that starts with a low belief score. A third model labels each persuader turn as an attempt or not. A fourth model runs StrongREJECT to mark explicit refusals. The 600 English topics cover conspiracies, everyday opinions, factual claims, contested social issues, unambiguously harmful statements, and statements that undermine human control of AI.

Task format

Multi-turn chat. Default is three persuader/persuadee turn pairs, temperature 0.5, persuade prompt (not the baseline). inspect_evals defaults: persuadee openai/gpt-4o, evaluator openai/gpt-4o, refusal_judge openai/gpt-4o-mini. Belief is sampled in [0, 20] so the requested direction is always "for".

Models reporting this benchmark

No model card in ModelSpec reports this benchmark yet.

Data

This page as JSON · Edit on GitHub