WildHallucinations

What it measures

Evaluates open-ended factual generation on long-tail entities mined from real user WildChat logs.

Reading the measurement

Accuracy is only part of the picture. Inspect unsupported answers, abstentions, language coverage and the evidence available to the model. A confident response can still be wrong, and a broad average can hide a specialist gap.

Published evaluation

Fraction of generated atomic claims supported by the evaluation pipeline over all entities. Abstention and entity difficulty affect interpretation.

Where this fits in an evaluation suite

Read accuracy together with error and abstention rates. Domain averages can hide language gaps, specialist failures, and poorly calibrated confidence.

Start with a capability hypothesis, then choose a second test with a different failure mode. Keep the checkpoint, reasoning effort, agent tools, sampling policy, and scoring procedure fixed when measuring a training intervention. Record failures and missing outputs in the denominator.

A reproducible comparison

Fraction of generated atomic claims supported by the evaluation pipeline over all entities. Abstention and entity difficulty affect interpretation.

Publish the exact run configuration, item coverage, score units, and available uncertainty. If a source supplies only an aggregate or a single run, preserve that limit. A new model name or a rounded score alone does not establish a capability improvement.