SpikingKiki-35B-A3B-V4

Lossless ANN-to-SNN (LAS) conversion of Qwen3.6-35B-A3B weights. Each original linear module replaced by a rate-coded integrate-and-fire surrogate (T=128, threshold=0.0625, tau=1.0).

Numbers

• 40 / 40 transformer blocks converted (Phase C)
• 31 070 per-module .npz weight files
• 99.5 % sane modules (validation Phase E, N=200)
• Built on Mac Studio M3 Ultra, 1 h 08 wall-clock via 4-worker block-range parallelism

Artifacts

• block_NN_MODULE.npz — per-module spiking weights
• lif_metadata.json — uniform LIF params
• No checkpoint bundling — 70 GB is easier to stream module-wise

Source / wiring

hypneum-lab/micro-kiki / scripts/convert_spikingkiki_35b.py. Consumed by the dream-of-kiki substrate under DREAM_MICRO_KIKI_REAL=1.

🇪🇺 EU AI Act transparency

This adapter is provided as a fine-tuned LoRA under the AI Act framework (Regulation EU 2024/1689). Compliance metadata:

Field	Value
Provider	L'Électron Rare (clemsail / electron-rare)
Role under AI Act	GPAI provider for this adapter
Base model	Qwen/Qwen3.6-35B-A3B — see upstream provenance
Adapter type	LoRA / PEFT — adapter weights only; base unchanged
Training data origin	L'Électron Rare proprietary technical corpus + curated public docs
License	Apache-2.0 (adapter). Upstream base licence applies separately.
Intended use	Multi-domain technical assistance — engineering, KiCad, embedded, code, FR/EN chat
Out of scope	Healthcare diagnosis, legal advice, autonomous safety-critical decisions, generation of malicious code
Risk classification	Limited risk — Article 50 transparency obligations apply
Copyright respect	Training data does not include scraped copyrighted material. Opt-out signals (robots.txt, ai.txt) are honoured for web-sourced data.
Full provenance	https://github.com/L-electron-Rare/eu-kiki/tree/main/docs/provenance
Contact	postmaster@saillant.cc — biased output reports, copyright concerns, etc.

⚠️ You are using an AI model. Outputs may be inaccurate, biased or fabricated. Do not act on them without independent verification, especially in regulated domains.