Single-answer Atomic Long-form Target (SALT) Benchmark

What is Single-answer Atomic Long-form Target (SALT)?

SALT is a procedurally generated benchmark for evaluating correctness and uncertainty estimates in long-form language-model outputs against deterministic…

Released

2026-07-04

Evaluates

General AI, Language & Knowledge, cs.AI

Openness

unknown

Importer

Claire Radar

Review status

ai-reviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2607.03870

Single-answer Atomic Long-form Target (SALT) paper and code

• Paper https://arxiv.org/abs/2607.03870
• Code https://github.com/IdoAmit198/SALT

Which sources cite Single-answer Atomic Long-form Target (SALT)?

• Single-answer Atomic Long-form Target (SALT)

Related General AI, Language & Knowledge, cs.AI 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