Corpus Agentis
The field book to agent ecosystems
The field book to agent ecosystems
Sources · T1

arXiv

Every datapoint in the Fieldbook that cites arXiv as a source, with the credibility tier it carries and a link to where the claim appears. The public provenance record for this source.

T1 · authoritative 40 of 97 shown
Cited by
Relationship (triple)Memory schemas
↗ arxiv.org · in the corpus
SummaryMemory schemas
↗ arxiv.org · in the corpus
Summarization / CompactionRetrieval patterns
↗ arxiv.org · in the corpus
Reflection / Self-EditRetrieval patterns
↗ arxiv.org · in the corpus
HyDE (Hypothetical Document Embedding)Retrieval patterns
↗ arxiv.org · in the corpus
RAG / Direct Vector SearchRetrieval patterns
↗ arxiv.org · in the corpus
Agent-Queried / Tool-Call RetrievalRetrieval patterns
↗ arxiv.org · in the corpus
Context Injection / Prompt StuffingRetrieval patterns
↗ arxiv.org · in the corpus
Temporal / Bi-Temporal QueryRetrieval patterns
↗ arxiv.org · in the corpus
LOKA ProtocolProtocols
↗ arxiv.org · in the corpus
Implicit toolbelt (ReAct)Ontologies
↗ arxiv.org · cited source
AgentDojoSafety evals
↗ arxiv.org · in the corpus
InjecAgentSafety evals
↗ arxiv.org · in the corpus
AgentHarmSafety evals
↗ arxiv.org · in the corpus
Agent Security Bench (ASB)Safety evals
↗ arxiv.org · in the corpus
CybenchSafety evals
↗ arxiv.org · in the corpus
CyberSecEval 3 (Purple Llama)Safety evals
↗ arxiv.org · in the corpus
τ-bench (tau-bench)Safety evals
↗ arxiv.org · in the corpus
JailbreakBenchSafety evals
↗ arxiv.org · in the corpus
HarmBenchSafety evals
↗ arxiv.org · in the corpus
StrongREJECTSafety evals
↗ arxiv.org · in the corpus
HackAPromptSafety evals
↗ arxiv.org · in the corpus
AdvBench (GCG attack)Safety evals
↗ arxiv.org · in the corpus
SafetyBenchSafety evals
↗ arxiv.org · in the corpus
SHADE-ArenaSafety evals
↗ arxiv.org · in the corpus
τ²-Bench (tau-squared)Safety evals
↗ arxiv.org · in the corpus
BIG-Bench Hard (BBH)Benchmarks
↗ arxiv.org · in the corpus
RecursiveTopologies
↗ arxiv.org · in the corpus
Memory-firstTopologies
↗ arxiv.org · in the corpus
Self-Consistency (N-sample)Topologies
↗ arxiv.org · in the corpus
Self-Consistency (N-sample)Topologies
↗ arxiv.org · in the corpus
Fully-Connected DebateTopologies
↗ arxiv.org · in the corpus
Fully-Connected DebateTopologies
↗ arxiv.org · in the corpus
Judicial Committee (quorum)Topologies
↗ arxiv.org · in the corpus
Judicial Committee (quorum)Topologies
↗ arxiv.org · in the corpus
Adversarial ReviewTopologies
↗ arxiv.org · in the corpus
Field notes

arxiv.org · cited 97 citations across the Fieldbook

arXiv (arxiv.org) is a primary or authoritative reference: a vendor statement about its own product, or an established research body, standards organisation or government source, strong enough to underpin a confirmed datapoint. It is cited 97 times across the Fieldbook, backing datapoints in memory schemas, retrieval patterns, protocols, ontologies, safety evals, benchmarks and topologies. Every figure the corpus draws from this source is graded T1, so the credibility of the claim travels with the claim itself.

Show more

Every datapoint below links to where the claim appears in the corpus, where a dedicated view exists, and to the exact source URL it rests on. Any reader can follow a fact back to its origin and check it, rather than take it on trust.

From the corpus, curated by Brandon Chaplin