{"id":49061480,"url":"https://github.com/hypneum-lab/nerve-wml","last_synced_at":"2026-04-25T02:01:12.067Z","repository":{"id":352309334,"uuid":"1214356656","full_name":"hypneum-lab/nerve-wml","owner":"hypneum-lab","description":"Substrate-agnostic nerve protocol for inter-WML communication — discrete neuroletters, γ/θ multiplexing, sparse learned topology","archived":false,"fork":false,"pushed_at":"2026-04-23T23:44:36.000Z","size":22237,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"master","last_synced_at":"2026-04-24T01:38:21.778Z","etag":null,"topics":["continual-learning","multiplexing","neural-protocol","python","reproducible-research","spiking-neural-network","substrate-agnostic"],"latest_commit_sha":null,"homepage":null,"language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/hypneum-lab.png","metadata":{"files":{"readme":"README.md","changelog":"CHANGELOG.md","contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":"CITATION.cff","codeowners":null,"security":"SECURITY.md","support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":".zenodo.json","notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2026-04-18T13:14:11.000Z","updated_at":"2026-04-23T23:43:37.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/hypneum-lab/nerve-wml","commit_stats":null,"previous_names":["electron-rare/nerve-wml","genial-lab/nerve-wml","hypneum-lab/nerve-wml"],"tags_count":37,"template":false,"template_full_name":null,"purl":"pkg:github/hypneum-lab/nerve-wml","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hypneum-lab%2Fnerve-wml","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hypneum-lab%2Fnerve-wml/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hypneum-lab%2Fnerve-wml/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hypneum-lab%2Fnerve-wml/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/hypneum-lab","download_url":"https://codeload.github.com/hypneum-lab/nerve-wml/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/hypneum-lab%2Fnerve-wml/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":32247508,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-24T13:21:15.438Z","status":"online","status_checked_at":"2026-04-25T02:00:06.260Z","response_time":59,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["continual-learning","multiplexing","neural-protocol","python","reproducible-research","spiking-neural-network","substrate-agnostic"],"created_at":"2026-04-20T02:16:11.420Z","updated_at":"2026-04-25T02:01:12.029Z","avatar_url":"https://github.com/hypneum-lab.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# nerve-wml\n\n[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.19656342.svg)](https://doi.org/10.5281/zenodo.19656342)\n[![OSF](https://img.shields.io/badge/OSF-10.17605%2FOSF.IO%2FQ6JYN-lightgrey)](https://doi.org/10.17605/OSF.IO/Q6JYN)\n\n**Substrate-agnostic nerve protocol for inter-module communication in hybrid neural systems.**\n\nCitation : each release is archived on Zenodo (concept DOI [10.5281/zenodo.19656342](https://doi.org/10.5281/zenodo.19656342) resolves to the latest version) and linked to the parent programme's OSF pre-registration ([10.17605/OSF.IO/Q6JYN](https://doi.org/10.17605/OSF.IO/Q6JYN)).\n\nResearch engine that validates a discrete-code communication layer between heterogeneous neural modules (World Model Languages, or WMLs). Modules exchange **neuroletters** over a sparse learned topology, multiplexed on gamma/theta rhythms, and converted between local codebooks by per-edge transducers. The paper draft is at [`papers/paper1/main.tex`](papers/paper1/main.tex); the full spec is at [`docs/superpowers/specs/2026-04-18-nerve-wml-design.md`](docs/superpowers/specs/2026-04-18-nerve-wml-design.md).\n\n## Status — v1.8.0 (2026-04-24, on PyPI)\n\nInstallable via `pip install nerve-wml`. For the real\n`kiki_oniric.axioms` integration (dream-of-kiki bridge), side-install\n`dreamofkiki` first (PyPI rejects VCS URLs in published metadata, so\nno `[axioms]` extras group is shipped):\n\n```bash\npip install \"dreamofkiki @ git+https://github.com/hypneum-lab/dream-of-kiki@v0.9.1\"\npip install nerve-wml\n```\n\nSix releases landed on 2026-04-21 → 2026-04-24 (v1.4.0 → v1.8.0) on top\nof the v1.2.3 scientific baseline; see\n[§ Post-v1.2.3 API additions](#post-v123-api-additions-2026-04-21) below\nor [`CHANGELOG.md`](CHANGELOG.md) for the per-version diff. The **scientific\nclaims below are the v1.2.3 baseline and remain load-bearing** — the newer\nreleases added opt-in knobs (plasticity schedule, Gumbel-softmax gating,\nspectrogram encoder, dreamOfkiki axiom bridge scaffold) and the\n`nerve_wml.methodology` submodule with the four MI robustness primitives\n(null-model, bootstrap CI, Miller-Madow, Kraskov KSG, MINE) — all without\nchanging any headline measurement.\n\nThe project is empirically defensible across three experimental axes: real\ndata, architecture scale, and temporal streaming. Two claims are quantified:\n\n**Claim A — Substrate-agnostic polymorphism (task competence converges).**\nThree structurally distinct substrates (stateless MLP, spiking LIF with surrogate-gradient, attention-based Transformer) reach comparable accuracies via the shared Nerve Protocol.\n\n**Claim B — Substrate-agnostic information transmission (codes align).**\nIndependent substrates share 91–96 % of their emitted code information; a frozen LIF can recover a trained MLP's task competence via a learned linear transducer.\n\n### Headline measurements\n\n| Axis | Finding | Reference |\n|---|---|---|\n| **Pool scaling law** (MLP ↔ LIF, HardFlow) | $N=2 \\to 10.71\\%$, $N=16 \\to 6.71\\%$, $N=32 \\to 2.39\\%$, $N=64 \\to 2.73\\%$ plateau. 5 % contract holds distributionally at $N \\geq 32$. | `figures/w2_hard_scaling.pdf` |\n| **Triple-substrate pool** (MLP + LIF + TRF) | $N=15 \\to 8.16\\%$, $N=30 \\to 5.86\\%$, $N=60 \\to 4.33\\%$ | v1.1.4 |\n| **Mutual information** (codes MLP ↔ LIF) | $\\mathrm{MI}/H = 0.91$ at $N=1$ (5 seeds), **0.96** at $N=16$ pool (192 cross-pairs) | `figures/info_transmission.pdf` |\n| **Round-trip fidelity** (MLP → LIF → MLP) | **0.99** mean (3 seeds) | v0.8 |\n| **Cross-substrate merge** (LIF fed by MLP codes only) | **0.97** mean (3 seeds) | v0.8 |\n| **MNIST real data** | MLP 0.942, LIF 0.941, gap **1.03 %**, MI/H **0.882** | `figures/mnist_scaling.pdf` |\n| **MoonsTask** (2nd distribution) | MI/H = **0.74** (3 seeds) | v1.1.4 |\n| **Architecture scale** ($d_\\text{hidden}=128$) | Gap AMPLIFIES to 26 % on XOR (arch vs pool scale are orthogonal); Claim B survives | `figures/bigger_arch_scaling.pdf` |\n| **Temporal streaming** (16-token sequence) | MI/H = **0.72** at trained step, **0.71** at filler step — structural alignment | `figures/temporal_info_tx.pdf` |\n| **Platonic RH alignment** (Huh 2024, pre-VQ mutual-kNN) | MLP ↔ LIF = **0.174** at k=10 (18.8× random, 3 seeds); stable across k∈[5,50] | `figures/platonic_rh_alignment.json` |\n| **Real neural data** (Sleep-EDF EEG, v1.6.0) | See paper Test (9); Claim B confirmed on 5-class sleep-stage via `MlpWML.from_spectrogram` + d_hidden=128 | `figures/mi_eeg_d128_spectro.json` |\n| **Frozen-encoder baseline** (review F3, v1.7.0) | Shared MI/H=0.95 (matches nerve-wml Test 1), Distinct MI/H=0.76 (without shared frontend); Claim B reframed as \"VQ protocol supplies shared frontend through codebook\" | `figures/baseline_frozen_encoder.json` |\n| **Matched-capacity scale sweep** (Sleep-EDF, v1.7.0) | Sweet spot at d=128: MI/H=0.72, MLP=0.82, LIF=0.83, gap=0.006. Scale-invariant polymorphy at d ∈ {32, 64, 128}. d=16 insufficient for LIF convergence on real EEG; d=256 MLP overfits while LIF holds | `figures/eeg_matched_scale_sweep.json` |\n| **Direction stability** (LIF ≥ MLP on hard task) | **19/20** pairwise seeds (4/5 at N=2; 5/5 at N=16, 32, 64) + 5/5 triple-substrate, preserved on Sleep-EDF (+0.007 LIF edge) | — |\n\nLIF's spike dynamics give it a substrate-intrinsic $\\sim 2$–$3\\%$ expressivity edge on XOR-style boundaries (plateau floor). Pool averaging compresses this, architecture width amplifies it.\n\n### Seven concrete findings\n\n1. **The original 12.1 % gap was a decoder asymmetry bug, not a substrate limit.** LIF had a fixed cosine decoder, MLP had a learned head; symmetrizing flipped the sign (LIF now leads).\n2. **Single-seed measurements lie.** Multi-seed revealed the N=16 median is 6.7 %, not the lucky 1.68 %.\n3. **Scaling law is real and monotonic.** Four-point decay $10.7\\% \\to 6.7\\% \\to 2.4\\% \\to 2.7\\%$ plateau.\n4. **Claim B is empirical, not architectural.** MI 0.91–0.96, round-trip 0.99, cross-merge 0.97.\n5. **Substrate-direction is stable in 19/20 seeds.** LIF's spike edge is a real property, not noise.\n6. **Architecture scale and pool scale are orthogonal.** Pool compresses the gap; arch width amplifies it.\n7. **Code alignment is structural, not task-gated.** MI at filler timesteps $\\approx$ MI at trained timesteps (0.71 vs 0.72).\n\n### Methodological findings (v1.2.1–v1.2.3)\n\n- **MI/H vs CKA on the same argmax codes** (v1.2.1). Mean 0.953 (MI/H) vs 0.910 (CKA argmax one-hot) over 3 seeds. The 4.3 pp gap tracks soft many-to-one code mappings that kernel-alignment metrics miss. MI/H is not CKA renamed — it is the discrete-protocol cousin with measurably different semantics. See `scripts/measure_cka_vs_mi.py` and `docs/positioning.md`.\n- **Related Work verified** (v1.2.2). Paper §Related Work cites Kornblith 2019 CKA, Morcos 2018 PWCCA, Moschella 2022 relative representations (ICLR 2023), Saxe 2024 universality, and Hinton 2015 KD — all verified via WebFetch, provenance table in `docs/positioning.md`.\n- **KD match-compute ablation honest verdict** (v1.2.3). At matched compute on HardFlowProxyTask (3 seeds), cross-merge (0.508) ≈ KD-through-transducer (0.520) within noise. Vanilla Hinton KD (0.534) is best because the student can re-train its core. Cross-merge's contribution is **methodological, not performance-based**: it isolates protocol channel capacity from student learning capacity by freezing both substrates and supervising with ground-truth labels only. See `scripts/measure_kd_ablation.py`.\n\n### What the paper genuinely claims vs not\n\nThree findings probably novel: (1) the four-point scaling law with plateau at $\\sim 2\\text{–}3\\%$ substrate-intrinsic floor, (2) reproducible $\\sim 2\\text{–}3\\%$ LIF spike-expressivity edge over matched-capacity MLP on XOR-on-noise (19/20 seeds), (3) orthogonality of pool-scale (compresses gap) and architecture-scale (amplifies gap).\n\nThe paper explicitly does **not** claim: a new learning algorithm, superiority over knowledge distillation on task accuracy, or universal representations — that debate is addressed by Saxe 2024 and the Nature MI 2025 editorial (s42256-025-01139-y) cited in `docs/positioning.md`.\n\n### Cross-lab methodology commitment\n\nThe sister project `bouba_sens` (2026-04-21, `github.com/hypneum-lab/bouba_sens` tag `v0.5.0`) demonstrated that pre-registered findings in this programme must pass three critical tests before publication: **null-model partition controls**, **bootstrap confidence intervals** on sub-threshold effects, and **multi-estimator robustness** checks for MI-based claims. As of **v1.5.3** (2026-04-21) all three checks are implemented in `nerve_wml.methodology` and applied to the MI/H headline: **null-model rejects chance** at z \u003e 1000 (p \u003c 10⁻³ over 3 seeds × 1000 shuffles), **bootstrap** gives CI95 [0.82, 0.99] intra-seed width ~0.005, and **discrete cross-estimator robustness** holds between plug-in and Miller-Madow (Δ = 0.007). Two continuous estimators (Kraskov KSG and MINE) were applied to the pre-VQ embeddings; they diverge by an order of magnitude (KSG 0.09, MINE 0.99), making the pre-VQ absolute MI magnitude an open methodological question — see paper §Information Transmission Test (7). The post-VQ discrete MI/H headline is unaffected by this ambiguity.\n\n## Status — 11 gates\n\n| Tag | What it proves |\n|---|---|\n| [`gate-p-passed`](../../releases/tag/gate-p-passed) | Track-P protocol simulator correct on toy signals |\n| [`gate-w-passed`](../../releases/tag/gate-w-passed) | `MlpWML` and `LifWML` interoperate with \u003c 5 % gap through the same nerve (N=4) |\n| [`gate-m-passed`](../../releases/tag/gate-m-passed) | Merge fine-tunes only transducers; retains ≥ 95 % of mock-baseline accuracy |\n| [`gate-m2-passed`](../../releases/tag/gate-m2-passed) | Four scientific shortcuts from §13.1 resolved with honest measurements |\n| [`gate-scale-passed`](../../releases/tag/gate-scale-passed) | Polymorphie + continual learning hold at N=16 pools; router stays connected to N=32 |\n| [`gate-interp-passed`](../../releases/tag/gate-interp-passed) | Per-WML `code → concept` semantics table rendered as HTML |\n| [`gate-neuro-passed`](../../releases/tag/gate-neuro-passed) | LifWML → INT8 artefact → pure-numpy mock runner (Loihi / Akida stubs documented) |\n| [`gate-dream-passed`](../../releases/tag/gate-dream-passed) | ε-trace consolidation bridge to dream-of-kiki (schema v0; partial — awaits kiki_oniric v0.5+) |\n| [`gate-adaptive-passed`](../../releases/tag/gate-adaptive-passed) | Per-WML alphabet shrinks/grows via `active_mask` + transducer resize |\n| [`gate-llm-advisor-passed`](../../releases/tag/gate-llm-advisor-passed) | Env-gated, never-raising `NerveWmlAdvisor` for micro-kiki, \u003c 50 ms warm latency |\n\nPaper drafts: `paper-v0.2-draft` … `paper-v0.9-draft` track the iterations that produced the v1.2 claims above. Release tags `v1.0.0`, `v1.1.0` … `v1.1.4`, `v1.2.0`, `v1.2.3`, `v1.3.0`, `v1.4.0`, `v1.5.0`, `v1.5.1` archive the code snapshots; see [`CHANGELOG.md`](CHANGELOG.md) for per-version findings.\n\n## Post-v1.2.3 API additions (2026-04-21)\n\nThree issues filed by downstream consumers (`bouba_sens`, `dream-of-kiki`)\nlanded on 2026-04-21 as opt-in knobs — **no change to v1.2.3 headline\nnumbers**, all new behaviour is off by default.\n\n| Release | Issue | Feature | Motivation (downstream) |\n|---------|-------|---------|-------------------------|\n| **v1.4.0** | [#4](https://github.com/hypneum-lab/nerve-wml/issues/4) | `GammaThetaMultiplexer` gains `plasticity_schedule` + `constellation_lock_after` | `bouba_sens` B-1 Amedi-2007 gap directionally falsified in 4/5 worlds; biologically-distinct T1/T2 plasticity profiles are the probe. |\n| **v1.5.0** | [#5](https://github.com/hypneum-lab/nerve-wml/issues/5) | `Transducer` gains `TransducerGating.GUMBEL_SOFTMAX` (opt-in soft distribution) | `bouba_sens` B-2 Me3-delta under-threshold in 5/5 worlds; hard argmax gating may be too abrupt for post-lesion MI migration. |\n| **v1.5.0** | [#7](https://github.com/hypneum-lab/nerve-wml/issues/7) | `MlpWML.from_spectrogram(...)` factory + `SpectrogramEncoder` | DRY: `bouba_sens` MIT-BIH ECG fetcher + future Studyforrest audio share one canonical STFT → carrier path. |\n| **v1.5.0** | [#6](https://github.com/hypneum-lab/nerve-wml/issues/6) | `nerve_core.from_dream_of_kiki(...)` scaffold (runtime gated upstream) | Pin the public axiom-bridge contract today so `bouba_sens` can plumb the call site before `dream-of-kiki` publishes its versioned `axioms` API. |\n| **v1.5.1** | — | Packaging: `pyproject.toml` `version` sync (stale `1.4.0` on the v1.5.0 commit), `CITATION.cff` keeps concept DOI only. | v1.5.0 shipped with a stale version field — first PyPI release carries the correct metadata. |\n\nDesign docs: [`docs/integration-dream-of-kiki.md`](docs/integration-dream-of-kiki.md), changelog files at [`docs/changelog/v1.4.0.md`](docs/changelog/v1.4.0.md) and [`docs/changelog/v1.5.1.md`](docs/changelog/v1.5.1.md).\n\n## Install\n\n```bash\n# From PyPI (v1.5.1+)\npip install nerve-wml\n\n# From source, with dev extras (tests + lint)\ngit clone https://github.com/hypneum-lab/nerve-wml.git\ncd nerve-wml\nuv sync --all-extras\n```\n\nPython 3.12+, macOS arm64 (MLX-friendly) or Linux x86_64. No vendor SDK deps are pulled by default (Loihi, Akida, dream-of-kiki, sentence-transformers are all optional integrations).\n\n## Run the suite\n\n```bash\nuv run pytest -m \"not slow\"    # 220+ tests under 80 s on commodity M-series\nuv run pytest                  # full suite incl. paper figure rendering\nuv run pytest --cov=nerve_core --cov=track_p --cov=track_w --cov=bridge --cov=harness --cov=interpret --cov=neuromorphic\n```\n\n## Reproduce the gate numbers\n\n```bash\nuv run python scripts/track_p_pilot.py       # Gate P (+ Task 6 ablation)\nuv run python scripts/track_w_pilot.py       # Gate W\nuv run python scripts/track_w_pilot.py scale # Gate Scale (N=16, N=32)\nuv run python scripts/merge_pilot.py         # Gate M\nuv run python scripts/interpret_pilot.py     # Gate Interp (emits reports/interp/*.html)\nuv run python scripts/adaptive_pilot.py      # Gate Adaptive\n```\n\n## Reproduce the v1.1 / v1.2 findings\n\n```bash\n# v1.1 scaling law + information transmission + triple substrate\nuv run python scripts/render_scaling_figure.py      # 4-point pool scaling (N=2..64)\nuv run python scripts/render_info_tx_figure.py      # MI + round-trip + cross-merge\nuv run python scripts/measure_info_transmission.py  # full info-tx battery\n\n# v1.2 real data + bigger arch + temporal\nuv sync --extra mnist                               # pull torchvision\nuv run python scripts/render_mnist_figure.py        # MNIST Claims A + B\nuv run python scripts/render_bigger_arch_figure.py  # d=128 gap amplification\nuv run python scripts/render_temporal_figure.py     # streaming MI per timestep\n```\n\n## Build the paper\n\n```bash\nuv run python scripts/render_paper_figures.py   # regenerate figures from frozen golden NPZs\ncd papers/paper1 \u0026\u0026 tectonic main.tex           # or pdflatex, bibtex, pdflatex, pdflatex\n```\n\n## Integrations (env-gated, default off)\n\n- **Dream consolidation**: `DREAM_CONSOLIDATION_ENABLED=1` + install `dream-of-kiki` locally → `bridge.dream_bridge.DreamBridge`.\n- **LLM advisor (micro-kiki)**: `NERVE_WML_ENABLED=1` + `NERVE_WML_CHECKPOINT_PATH=/path/to/checkpoint` → `bridge.kiki_nerve_advisor.NerveWmlAdvisor`. Wiring recipe: [`docs/integration/micro-kiki-wiring.md`](docs/integration/micro-kiki-wiring.md).\n- **Neuromorphic hardware**: install `lava-nc` or `akida` → wire in `neuromorphic.loihi_stub` / `neuromorphic.akida_stub`. Schema v0: [`docs/neuromorphic/deployment-guide.md`](docs/neuromorphic/deployment-guide.md).\n\n## Cited in\n\n- **dreamOfkiki — Paper 1 v0.2 (2026-04-19), §7.4 cross-substrate portability** — [github.com/hypneum-lab/dream-of-kiki](https://github.com/hypneum-lab/dream-of-kiki). The Gate W and Gate M measurements reported here (MlpWML / LifWML polymorphism on FlowProxyTask and HardFlowProxyTask) provide the empirical corroboration cited in Paper 1 as independent evidence of the substrate-agnosticism principle (DR-3 Conformance Criterion). OSF pre-registration: [10.17605/OSF.IO/Q6JYN](https://doi.org/10.17605/OSF.IO/Q6JYN).\n\n## Program context\n\nThis repository is part of **hypneum-lab**, which develops executable formal frameworks for cognitive AI. The programmatic parent is `dreamOfkiki` (paper 1 formal framework, paper 2 empirical); `nerve-wml` is the reference implementation for the substrate-agnostic communication principle.\n\nSibling repositories:\n\n- [dream-of-kiki](https://github.com/hypneum-lab/dream-of-kiki) — formal framework (axioms DR-0..DR-4, Conformance Criterion, Paper 1)\n- [kiki-flow-research](https://github.com/hypneum-lab/kiki-flow-research) — Wasserstein-gradient-flow engine (upstream)\n- [micro-kiki](https://github.com/hypneum-lab/micro-kiki) — 35 domain-expert MoE-LoRA deployable instance (advisor consumer)\n- **nerve-wml** (this repo) — substrate-agnostic nerve protocol + cross-substrate polymorphism\n\n## Repository layout\n\n```\nnerve_core/        Neuroletter, Nerve + WML Protocols, invariants (N-1..N-5, W-1..W-4)\ntrack_p/           Track-P — SimNerve, VQCodebook, Transducer, SparseRouter, AdaptiveCodebook\ntrack_w/           Track-W — MockNerve, MlpWML, LifWML, toy tasks, training loop, pool factory\nbridge/            Merge, dream, LLM advisor — SimNerveAdapter, MergeTrainer, DreamBridge, NerveWmlAdvisor\nharness/           R1 reproducibility — run_registry\ninterpret/         Gate Interp — code_semantics, clustering, HTML renderer\nneuromorphic/      Gate Neuro — spike_encoder, INT8 export, mock_runner, vendor stubs\nscripts/           All gate pilots + figure renderers + freeze_golden\ntests/             Unit + integration + golden NPZ regressions\ndocs/              specs/, integration/, neuromorphic/, dream/, interpret/\npapers/paper1/     LaTeX source + bib + Makefile (figures regenerated deterministically)\n```\n\n## License\n\nMIT (code) + CC-BY-4.0 (docs).\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fhypneum-lab%2Fnerve-wml","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fhypneum-lab%2Fnerve-wml","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fhypneum-lab%2Fnerve-wml/lists"}