Model techniques map
Techniquesevaluationbenchmark

specific method · filed under evaluation

PaperBench

A paper-replication evaluation using hierarchical rubrics that decompose each task into smaller subtasks with grading criteria.

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.

For objective evaluation, we develop rubrics that hierarchically decompose each replication task into smaller sub-tasks with clear grading criteria.

usedevaluation onlyin gpt-oss-120b and gpt-oss-20bOpenAI

Filed alongside

Other methods under evaluation :: benchmark.