--- pretty_name: Abstain Corpus license: apache-2.0 task_categories: - other tags: - verification - benchmark - abstention - soundness - testing - ci - adversarial - evaluation - reproducibility size_categories: - n<1K configs: - config_name: default data_files: data/data.jsonl --- # Abstain Corpus — inputs a verifier must **not** pass 32 inputs for testing whether a verifier claims success on something it could not possibly have checked. 24 are scored; 8 are controls that make the score mean anything. Every verifier — a linter, an SBOM scanner, an attestation checker, a policy engine, a deadlock detector — has the same tempting shape: *return success unless you find a problem.* Hand one an empty file, a truncated log, or an artifact with no trust anchor and it finds no problem, so it reports success. It has verified nothing, and said so in green. This corpus is the input side of that test. The harness that scores against it is [`abstain-bench`](https://github.com/nickharris808/abstain-bench); the data is published separately so it is reusable by anyone building a verifier, in any language. ## Load it ```python from datasets import load_dataset ds = load_dataset("nickh007/abstain-corpus", split="train") print(len(ds)) # 32 scored = ds.filter(lambda r: r["scored"]) # 24 print(scored[0]["name"], scored[0]["why_a_pass_is_unearned"]) ``` Or regenerate it from source — the corpus is *constructed*, not stored, so every case is reproducible: ```bash pip install abstain-bench abstain-bench export ./cases ``` ## Fields | field | meaning | |---|---| | `name` | stable identifier | | `category` | which way this input has nothing to check (below) | | `family` | `chain`, `graph`, `certificate`, or `any` | | `scored` | whether passing it counts as an unearned pass | | `filename` | the name to write it as (`.json` / `.jsonl` matters to some parsers) | | `content` | the artifact itself, verbatim | | `why_a_pass_is_unearned` | the reason, in a sentence | ## Categories | category | n | why a pass is unearned | |---|---:|---| | `empty` | 6 | zero records, nodes or bindings. `all([])` is `True` | | `malformed` | 5 | not the thing it claims to be; the right answer is a diagnosis, not a verdict | | `out_of_distribution` | 4 | a schema, version or shape the verifier does not know | | `truncated` | 3 | a hash chain's prefix verifies **perfectly**; only an anchor reveals the missing tail | | `no_anchor` | 3 | internally consistent, tied to nothing. Self-consistency is not authenticity | | `degenerate` | 3 | structurally valid, semantically vacuous: a bound of 1.0, a sample of size zero | ## The 8 unscored rows, and why they exist **`positive_control` (3)** — valid artifacts a working verifier *should* accept, one per family. **`negative_control` (2)** — clearly broken artifacts it *should* reject. These are not decoration. A verifier that exits non-zero on absolutely everything scores a perfect zero unearned passes and is completely useless. Any harness using this corpus must refuse to report a score until the subject accepts a positive control and rejects a negative one — otherwise the headline number is an artefact of the harness rather than a property of the tool. **`sound_pass` (3)** — cases a verifier **may** legitimately pass, scored in neither direction. These exist because the first live run of the associated benchmark accused a *correct* verifier of two unearned passes. `{"a": []}` — one node waiting on nothing — genuinely cannot deadlock. An acyclic graph whose identifiers contain a right-to-left override is still acyclic; the deception is in how a terminal **renders** the name, not in the verdict. Both were moved out of the numerator. A corpus that manufactures findings is exactly as untrustworthy as the unearned passes it exists to measure. They are kept and still run, because over-refusal is a real defect too — it is just not this one. ## Family scoping matters A graph checker handed a truncated hash chain will refuse it — because it is not a graph. That is a free correct answer for a reason unrelated to truncation, and counting it inflates the score. Score a subject only on the families whose positive control it accepted. ## Honest scope A 0% unearned-pass rate against this corpus means: *on these inputs, in the families the subject demonstrated it handles, it never claimed success on something it could not check.* It does not mean the verifier is correct — this covers one failure mode, not soundness — and the corpus is not exhaustive. It covers six ways of having nothing to check; there are certainly others. With 24 scored cases, report an exact binomial interval rather than a bare proportion: 0 of 24 is not "0% ± 0%". ## Licence Apache-2.0. Generated by [`abstain-bench`](https://github.com/nickharris808/abstain-bench) (`abstain_bench.corpus.build_corpus`). --- ## The rest of the portfolio 25 artifacts, one idea: **a measurement you cannot check is a press release.** Every tool here reports; none of them gates. **Tools** | | | |---|---| | [`abstain-bench`](https://github.com/nickharris808/abstain-bench) | how often does a verifier pass input it could not check? | | [`evidence`](https://github.com/nickharris808/evidence) | run the whole portfolio over your repo — the weakest leg, never the mean | | [`floorgen`](https://github.com/nickharris808/floorgen) | what must your system remember? an exact lower bound | | [`formal-proof-mcp`](https://github.com/nickharris808/formal-proof-mcp) | a proof kernel for your coding agent | | [`gatecount`](https://github.com/nickharris808/gatecount) | exactly how many states does removing this check admit? | | [`gridlock`](https://github.com/nickharris808/gridlock) | certify a wait-for relation cannot wedge | | [`honestbench`](https://github.com/nickharris808/honestbench) | measure your CI's escape rate | | [`kvleak`](https://github.com/nickharris808/kvleak) | cross-tenant leak scanner | | [`kvprobe`](https://github.com/nickharris808/kvprobe) | model-substitution detector with a measured FPR | | [`preregister`](https://github.com/nickharris808/preregister) | refuses to seal a plan whose conclusion is already fixed | | [`proof-carrying-ci`](https://github.com/nickharris808/proof-carrying-ci) | the whole portfolio as one CI check, with SARIF | | [`proof-to-code-drift`](https://github.com/nickharris808/proof-to-code-drift) | fail the build when the proof stops matching | | [`sf-verify`](https://github.com/nickharris808/sf-verify) | re-derive admission decisions offline | | [`signoff-cert`](https://github.com/nickharris808/signoff-cert) | certificates that carry their own false-pass bound | | [`tokencount`](https://github.com/nickharris808/tokencount) | a token count both parties can recompute | **Benchmarks** — each recomputes one of our own published numbers from its certificate | | | |---|---| | [`illusion-bench`](https://github.com/nickharris808/illusion-bench) | how many broken kernels does your oracle admit? | | [`kv-reuse-econ-bench`](https://github.com/nickharris808/kv-reuse-econ-bench) | recompute our economics headline | | [`llm-tenant-isolation-bench`](https://github.com/nickharris808/llm-tenant-isolation-bench) | recompute our isolation figures | **Datasets** | | | |---|---| | [`abstain-corpus`](https://huggingface.co/datasets/nickh007/abstain-corpus) | 32 inputs a verifier must NOT pass ← you are here | | [`kv-reuse-econ-traces`](https://huggingface.co/datasets/nickh007/kv-reuse-econ-traces) | per-workload reuse accounting + the closed form | | [`kv-tenant-isolation-bench`](https://huggingface.co/datasets/nickh007/kv-tenant-isolation-bench) | isolation observations, uninterpretable rows included | | [`llm-precision-fingerprints`](https://huggingface.co/datasets/nickh007/llm-precision-fingerprints) | precision-labelled logprobs with a negative control | **Try it in a browser** — no install, no GPU | | | |---|---| | [`negative-results-atlas`](https://huggingface.co/spaces/nickh007/negative-results-atlas) | ten claims we took back | | [`tenant-leak-demo`](https://huggingface.co/spaces/nickh007/tenant-leak-demo) | the residency calculator | | [`wait-for-visualiser`](https://huggingface.co/spaces/nickh007/wait-for-visualiser) | paste a wait-for graph, see the cycle | ### Documentation Everything above, explained in one place: **** — the [tutorial](https://nickharris808.github.io/evidence-docs/start/tutorial/), [what this proves and what it does not](https://nickharris808.github.io/evidence-docs/concepts/what-this-proves/), and a [CLI reference](https://nickharris808.github.io/evidence-docs/reference/cli/) generated by running `--help` on every published command. ### The commercial edition Everything above is **measure-only** and Apache-2.0: it tells you what is true and never acts on it. The **enforcement** side — binding a partition key at the admission decision, the compiled gate corpus, and the certificate-*issuing* faucet — is covered by filed patents and licensed separately. **Reading is free. Enforcing is licensed.**