conformance
x402-bazaar-conformance
x402 Bazaar conformance census
In one line: One probe per host listed in the public x402 Bazaars (Coinbase CDP, PayAI): does the payment door the index advertises answer the x402 protocol? For builders of agents that pay per call, and for x402 sellers checking their own door.
Use it
from datasets import load_dataset
ds = load_dataset("csoai/x402-bazaar-conformance", split="train")
print(ds[0])
Look up the signed measurements held about one x402 endpoint (paste… See the full description on the dataset page: https://huggingface.co/datasets/csoai/x402-bazaar-conformance.sparam-conformance
sparam-conformance
📖 Documentation site — the portfolio narrative, the concepts, a full walkthrough, and what all of this proves (and does not).
A labelled corpus of S-parameter networks with ground-truth physical verdicts —
and a scorer that grades any checker against it.
Why this exists
There is no public dataset of physically invalid S-parameter files. Everyone
building an RF validation tool tests it on files that happen to be lying around,
which means… See the full description on the dataset page: https://huggingface.co/datasets/nickh007/sparam-conformance.cis2-conformance
CIS-2 — conformance vectors for bit-identical fp32 transformer inference
Floating-point transformer inference is usually treated as unavoidably
nondeterministic across hardware. Reduction order, FMA contraction, denormal
handling and platform math libraries all differ between x86_64 and aarch64, and
between compilers, so "the same model on the same input" in practice means
"agrees to within a tolerance", not bit-for-bit.
CIS-2 is a written specification that removes those… See the full description on the dataset page: https://huggingface.co/datasets/aefinityAIINC/cis2-conformance.tokenizer-conformance
Tokenizer conformance fixtures
Reference inputs and Python fast-tokenizer outputs for tokenizer implementations.
The initial corpus contains 83 inputs in 30 categories, with 498 reference encodings across six tokenizers.
This is a regression dataset, not a model-quality benchmark.
Provenance and attribution
The input corpus and reference entries come from apocryphx's swift-transformers PR #360,
at commit ce847085784bacd8c3c15180c976b17c8ce73e31.
The corpus… See the full description on the dataset page: https://huggingface.co/datasets/pcuenq/tokenizer-conformance.dcp-conformance-corpus
DCP v1 Conformance Corpus
The official, language-neutral conformance test corpus of the Development Coordination Protocol (DCP) — an open, vendor-neutral vocabulary for exchanging project-coordination events (a task completed, a decision recorded, a finding raised…) between AI agents and humans.
A DCP implementation, in any language, is conformant when it accepts every accept case and rejects every reject case in this corpus, validating each payload as a DcpMessage against the… See the full description on the dataset page: https://huggingface.co/datasets/DevCoPro/dcp-conformance-corpus.JEP-Core-0.7-Conformance-Suite
JEP-Core 0.7 Conformance Suite
Companion test fixtures for draft-wang-jep-judgment-event-protocol-07.
Status
This repository is an implementation resource for JEP-Core 0.7. The current
Internet-Draft is the controlling specification. If a fixture, schema, or
validator here conflicts with the draft, the draft controls.
Canonical draft:
https://datatracker.ietf.org/doc/html/draft-wang-jep-judgment-event-protocol-07
Included
cases.jsonl — Dataset… See the full description on the dataset page: https://huggingface.co/datasets/hjs-spec/JEP-Core-0.7-Conformance-Suite.