PhantomFill Benchmark

What is PhantomFill?

PhantomFill measures schema-coerced fabrication in language models by asking questions on unanswerable inputs under three output formats: free text, JSON with…

Released

2026-06-11

Evaluates

General AI, Language & Knowledge, cs.LG

Openness

unknown

Importer

Claire Radar

Review status

unreviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2607.20492

PhantomFill paper and code

• Paper https://arxiv.org/abs/2607.20492
• Code https://github.com/ranausmanai/phantomfill

Which sources cite PhantomFill?

• PhantomFill

Related General AI, Language & Knowledge, cs.LG benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 3DCodeBench
• 5-Dialects-BN
• A Large-Scale Dataset and Benchmark
• A3S-Bench
• AARRI-Bench
• ABC-Bench
• ActBench