Dontopedia

PRBench

From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-06.)

PRBench has 21 facts recorded in Dontopedia across 2 references.

21 facts·20 predicates·2 sources

Mostly:sourced from num(1), comprises(1), comprises num(1)

Maturity scale raw canonical shape-checked rule-derived certified

Inbound mentions (1)

Other subjects in dontopedia point AT this entity as a value. These are inverse relationships — e.g. "X motherOf this subject" — and answer questions the forward facts can't. Grouped by predicate.

discussesDiscusses(1)

Other facts (20)

The long tail: predicates that appear too rarely to warrant their own section. Filter or scroll to find a specific one. Each row links to its source.

20 facts
PredicateValueRef
Sourced From Num>20[1]
ComprisesExpert Curated Tasks[1]
Comprises Num30[1]
Covers11 Subfields of Physics[1]
DesignsMulti Agent Evaluation Pipeline[1]
Evaluates End to End Reproduction ofComputational Results From Published Physics Papers[1]
Spans Num11[1]
Introduced inPrbench Paper[1]
SpansSubfields of Physics[1]
Is Benchmark forAI Agents[1]
Builds onAgentified Evaluation Paradigm[1]
Rdf:typeSoftware Benchmark[2]
Purposeevaluating AI agents on the end-to-end reproduction of computational results from published physics papers[2]
Number of Tasks30[2]
Number of Subfields11[2]
Sourced From InstitutionPeking University School of Physics[2]
Number of Research Groups20[2]
Uses Paradigmagentified evaluation paradigm[2]
Involves Pipelinemulti-agent evaluation pipeline[2]
Execution Environmentsandboxed environment[2]

Timeline

Timeline axis is valid_time — when each source says the fact was true in the world, not when Dontopedia learned about it. Retracted rows are kept for provenance; coloured stripes indicate the context kind.

sourcedFromNumblah/papers/part-17
>20
comprisesblah/papers/part-17
ex:expert-curated-tasks
comprisesNumblah/papers/part-17
30
coversblah/papers/part-17
ex:11-subfields-of-physics
designsblah/papers/part-17
ex:multi-agent-evaluation-pipeline
evaluatesEndToEndReproductionOfblah/papers/part-17
ex:computational-results-from-published-physics-papers
spansNumblah/papers/part-17
11
introducedInblah/papers/part-17
ex:prbench-paper
spansblah/papers/part-17
ex:subfields-of-physics
isBenchmarkForblah/papers/part-17
ex:ai-agents
buildsOnblah/papers/part-17
ex:agentified-evaluation-paradigm
typeblah/papers/17
ex:SoftwareBenchmark
labelblah/papers/17
PRBench
purposeblah/papers/17
evaluating AI agents on the end-to-end reproduction of computational results from published physics papers
numberOfTasksblah/papers/17
30
numberOfSubfieldsblah/papers/17
11
sourcedFromInstitutionblah/papers/17
ex:peking-university-school-of-physics
numberOfResearchGroupsblah/papers/17
20
usesParadigmblah/papers/17
agentified evaluation paradigm
involvesPipelineblah/papers/17
multi-agent evaluation pipeline
executionEnvironmentblah/papers/17
sandboxed environment

References (2)

2 references
  1. [1]Part 1711 facts
    ctx:discord/blah/papers/part-17
  2. [2]1710 facts
    ctx:discord/blah/papers/17
    • full textpapers-17
      text/plain2 KBdoc:agent/papers-17/769f76a6-a1b1-41a9-8172-fe068b039ca2
      Show excerpt
      [2026-04-19 10:01] lisamegawatts: https://megawatts.notion.site/Concepts-75967df4f4594361b0f6a8781bca895f?pvs=73 [2026-04-20 11:00] foxhop.: https://lumbda.com/ [2026-04-20 22:09] lisamegawatts: https://arxiv.org/pdf/2601.06793 CliffordNet:

See also

Keep researching

Missing something or suspicious of what's here? Kick off a research session — a Claude agent will investigate, cite its sources, and file new facts into a dedicated context you can review before accepting into the shared view.