specific method · filed under evaluation
Atomic binary rubric evaluation
This evaluation method scores task completion with atomic, binary rubric items, including code-based checks for deterministic properties.
- source
- 1
- models
- 2
- 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.
Task completion is evaluated using atomic, binary rubric items: code-based checks verify deterministic properties
usedevaluation onlyin MiMo-V2.6 general-agent task pipelineXiaomi
Filed alongside
Other methods under evaluation :: evaluation harness.
DeepSeek Harness Minimal modeHarborNeMo Evaluator SDKAutomatic in-flight checkpoint evaluation schedulingBenchmark evaluation frameworkCross-scaffold evaluationInline training evaluatorsIsolated container per rolloutJSON-structured answer-format promptmini-swe-agent harnessNeMo Gym and NeMo Evaluator-based harnessNeMo SkillsPool