ambiguous · filed under evaluation
ScaleAI Multi Challenge Multi Turn Instruction Following
A named multi-turn instruction-following benchmark; the evidence identifies it as an evaluation with an open-source packaged container but does not describe its tasks.
- source
- 1
- model
- 1
- lab adopt it
- 1
- strongest
- used
How sources treat it
One count per evidence span, weakest treatment to strongest.
used 1
Documented in
Evidence
1 span quoted from the sources, strongest treatment first.
the evaluations also used dedicated open-source packaged containers for ScaleAI Multi Challenge Multi Turn Instruction Following and KernelBench
usedevaluation onlyin Nemotron 3 UltraNVIDIA
Filed alongside
Other methods under evaluation :: benchmark.
High-confidence ProgramBench subsetArtificial Analysis Intelligence IndexBrowseComp with SearchCAD visual reproductionCapture-the-Flag (CTF) evaluationComputer use closed-loop evaluationDisabling tool searchDomain whitelistExecution-grounded evaluationF2P/P2F acceptance criterionFail-to-pass and pass-to-pass evaluation pointsGPU kernel optimization task suiteKernelBenchLiving in-house benchmark suiteLongBench v2Multiple-rollout evaluationPaperBenchPinchBenchProfBench with SearchSQuADSWE-bench VerifiedThree-configuration cyber range testingVals.aiVerified CUDA kernels