CogniRank: independent reproduction

This dataset repository distributes evaluation software, including a single self-contained cognirank.py. The script embeds the evaluation code, SmolCUA questions and answer keys, source checksums, and selected Procedural Pile row indices. Internet access is needed for Python packages, model weights, and the other pinned datasets.

Run

Use Linux, Python 3.11, git, a C/C++ build toolchain, and an NVIDIA GPU supporting CUDA and bfloat16, with enough memory for your model and the fixed batch sizes. Download cognirank.py from this repository using an account granted access. For example, after authenticating with Hugging Face:

hf download dotlabs/cognirank cognirank.py --repo-type dataset --local-dir .
python3.11 -m venv .venv
. .venv/bin/activate
python -m pip install --upgrade pip
python cognirank.py --verify
python cognirank.py --install-deps
python cognirank.py --model dotlabs/void.1 --output-dir void1-results

Multiple models run sequentially:

python cognirank.py --models dotlabs/void.1 MinimaLabs/KLM-1.5 LiquidAI/LFM3-1.7B --output-dir comparison
python cognirank.py --models dotlabs/void.1 MinimaLabs/KLM-1.5 LiquidAI/LFM3-1.7B --output-dir comparison --resume

--dry-run prints the protocol and model revisions without installing dependencies or running inference. --extract-only ./cognirank-source extracts the complete bundled source for inspection. --print-requirements prints dependencies. Installation explicitly executes pip in the current environment; inference permits Hugging Face custom model code, as the original evaluator does.

Known model revisions are pinned in source_manifest.json. Other model IDs resolve main once and record the immutable revision. To choose a different revision for one model, use --model-revision COMMIT_SHA. Use a new output directory for a different protocol, revision, or context. Resume preserves the resolved revision and verifies completed component checksums; SmolCUA resumes individual predictions, and Procedural Pile can re-score saved generations. Interrupted MARC/TTM components restart.

Exact protocol

Component	Examples	Weight	Metric
SmolCUA v1.4	1,200: three fixed passes of 400	30/95	Mean weighted pass score, 0–100
MARC	320: 4 families × 4 levels × 20	25/95	Mean exact-match accuracy of MARC-0…3
Procedural Pile	Fixed 350 across all 55 task families	25/95	Mean official verifier reward × 100
Tiny Theory of Mind	All 2,000	15/95	Token-normalized likelihood multiple-choice accuracy

CogniScore = (30*SmolCUA + 25*MARC + 25*ProceduralPile + 15*TTM)/95. All components are required; failures do not produce a partial composite. The rounded display weights are 31.58%, 26.32%, 26.32%, 15.79%; computation uses the exact fractions above.

Generation is greedy, bfloat16, CUDA device 0. SmolCUA uses automatic chat/plain formatting, batch 32, a 512-token response cap (48 for short formats). MARC uses the upstream plain demonstration prompts, deterministic seed string 0, batch 32, and 200 generated tokens. Procedural Pile uses automatic chat formatting with thinking disabled, batch 8, and 512 generated tokens. TTM uses plain continuations and lm-eval==0.4.12, batch 8. Full implementations are embedded; prompt construction, whitespace-aware scoring, and selection indices can be inspected with --extract-only.

Default requested evaluation context is 4,096. Extending a model's declared context is recorded as a context-extension experiment. Known fixed learned-position architectures are capped at their hard positional limit. Response budgets shrink to remaining space; input prompts are never truncated. A prompt that cannot fit fails explicitly. --eval-context 0 uses the native context, capped at 4,096 by this runner; changing context changes the experiment. Effective limits and shortened response budgets are recorded in component outputs.

Procedural Pile uses the pinned official verifier with inverse-math-normalize-progress-v1: it rejects a non-progressing malformed LaTeX fraction instead of hanging indefinitely. Other verifier timeouts fail the evaluation rather than silently assigning zero. The patch and upstream license are embedded in inverse_math_normalizer.py.

Immutable sources

• SmolCUA: dotlabs/smolcua, revision faf2b33f893573e23666a4e59c99a09d0b73ba51, runs/v1_4_seed_20261023/. Bundled questions SHA256: 74fb0a12a8dd75cbd8675d76e8bcf4a16e1b424668a195f4ff9a31da2becf065.
• MARC generator: FlameF0X/Micro-ARC/marc.py, revision 4ed8c708aa3e9665a9d9c81f2e3bc586cf60e14e.
• Procedural Pile: reasoning-core/procedural-pile, revision 0e3216537b0cb7df8709c6cae34bfceae6450068. Full test shard SHA256: 4ed623b9ad54a2f08483c8281825000e2f46c1e8ccfdeddd1db1cb1ca5c6cac8. The 350 selected IDs and family counts are embedded in source_manifest.json; selection version procedural-pile-350-v1, subset SHA256 67647511ff184a44c9f1cb7abaed66e0cd113f2deecafba95250b9a7338ddf6f. Twenty families contribute seven rows and 35 contribute six; exact hash-based selection is in the bundled parser.
• Official verifier: sileod/reasoning_core, revision 68b0b3a0f29128d2c56b3eafcdaa34777b9e1279.
• TTM: AxiomicLabs/Tiny_Theory_of_Mind, revision ba4c644ea7ace67e096305fe2c9f87c4cf4879ad, SHA256 58827a0361597ae4d5ffa79ddbab0936d1509cafcfc02cbb731e4d7a7f3584b8.

The source manifest includes the six example model repository names and exact revisions, every bundled source checksum, and source download URLs. Save the Hugging Face commit of this release and SHA256SUMS alongside your results.

Outputs and reproducibility limits

Each model directory contains run.json, results.json, per-component logs and predictions, environment.json, and pip-freeze.txt. comparison.json contains the completed model scores. The release identifier hashes bundled source files; resume rejects a different release. The single-file archive is checksum-verified before extraction.

Source datasets retain their upstream licenses and attribution. SmolCUA's source README and the inverse-math patch's MIT notice are included in the archive.