ORT (Out-of-Distribution Robustness Testing)

ORT checks whether an unlearning method erases knowledge of a real person across four query forms, not only the form used in the unlearning samples.

Also known as: ORT, ORT Benchmark, Out-of-Distribution Robustness Testing, LLM Unlearning Should Be Form-Independent

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
Subcategoryform-independent LLM unlearning (forget vs retain, four query formats)
Page statusactive
Metricjoint probability of the gold answer given the prompt
Directionlower_is_better
Dataset size35052
PublisherBeijing University of Posts and Telecommunications, Shandong University, and Institute of Automation, Chinese Academy of Sciences

What it measures

ORT is a text-only English unlearning test. Each target is a well-known real person. After unlearning, the model is queried on a forget set about that person and on a retain set about other people. The same facts appear in four formats: simple QA, fill-in-the-blank, multiple-choice labels, and subtoken-inducing QA that forces character-level tokens. A method that only suppresses the training form can still answer the other three. The paper calls that failure Form-Dependent Bias. Utility is checked separately on MMLU, TruthfulQA, TriviaQA, and AlpacaEval rather than inside ORT itself.

Task format

Single-target unlearning, then prompt-based probes. Training corpora come in three styles (unstructured text, refusal QA, preference pairs). Evaluation prompts are QA, fill-in-the-blank, multiple-choice, or subtoken QA on both the forget set and the retain set.

Models reporting this benchmark

No model card in ModelSpec reports this benchmark yet.

Data

This page as JSON · Edit on GitHub