Surgical Behavioral & Config Patches for Open-Weight Language Models

Stop re-downloading 16 GB full-model forks for a 64-column MLP fix or broken generation_config.json. Apply cryptographically verified .wsp v2.1 Copy-on-Write patches in under 5 seconds, or attest your own LoRA, merge, or fine-tune with WSP Verify.

$ pip install wsphub && wsp apply Qwen/Qwen3.5-0.8B --patch wsp-qwen3.5-0.8b-math-loop-exp8 --out ./patched

Verified Patch Catalog (P0 Config & P1 Weights)

Every patch is pre-registered, evaluated across 10,000 paired bootstrap and permutation tests, checked against the secret perturbed WSP-Holdout split and StrongREJECT/XSTest safety gates, and signed with Ed25519.

P0_CONFIG wsp-qwen2.5-coder-1.5b-rep-penalty-syntax-fix Target: Qwen/Qwen2.5-Coder-1.5B-Instruct

Qwen2.5-Coder-1.5B Repetition-Penalty Identifier & JSON Brace Fix

Neutralizes repetition_penalty=1.1 in Qwen/Qwen2.5-Coder-1.5B-Instruct generation_config.json that corrupts repeated variable names and nested closing braces in structured code/JSON output.

$ pip install wsphub && wsp apply Qwen/Qwen2.5-Coder-1.5B-Instruct --patch wsp-qwen2.5-coder-1.5b-rep-penalty-syntax-fix --out ./patched
Ed25519 Attestation: 95432ddaf5d64f26... Downloads: 0 View Full Verification →
P1_WEIGHTS wsp-qwen3.5-0.8b-math-loop-exp8 Target: Qwen/Qwen3.5-0.8B

Qwen3.5-0.8B 32k Thinking-Loop Surgical Cure (MIT CSAIL Exp-8)

Eliminates infinite 32,768-token chain-of-thought looping in Qwen/Qwen3.5-0.8B by modifying 64 output columns of mlp.down_proj in layers 15 and 19 (131,072 bf16 weights, 0.016% of parameters). Lifts independent MATH-500 accuracy from 5.0% (7.9% paper) to 51.0% (48.1% paper).

$ pip install wsphub && wsp apply Qwen/Qwen3.5-0.8B --patch wsp-qwen3.5-0.8b-math-loop-exp8 --out ./patched
Ed25519 Attestation: 243b0b7198656341... Downloads: 0 View Full Verification →
P0_CONFIG wsp-qwen3.5-0.8b-missing-genconfig-eos-fix Target: Qwen/Qwen3.5-0.8B

Qwen3.5-0.8B Missing generation_config & Dual Stop-Token Fix

Adds the missing generation_config.json to Qwen/Qwen3.5-0.8B with dual EOS stop tokens [248046 (<|im_end|>), 248044 (<|endoftext|>)] and verified sampling parameters, curing post-turn hallucination bleed.

$ pip install wsphub && wsp apply Qwen/Qwen3.5-0.8B --patch wsp-qwen3.5-0.8b-missing-genconfig-eos-fix --out ./patched
Ed25519 Attestation: 1a982181142ea44d... Downloads: 0 View Full Verification →

WSP Verify CI/CD & Supernova Settlement Rail

Automated regression testing and attestation for model builders, enterprise publishers, and AI coding agents via Supernova async escrow:

  • $0.25 — Remote Config Triage (Synchronous zero-GPU diagnostic in <1s)
  • $4.90 — Live GPU Diagnostic Sweep (Stop-token & sampling grid evaluation)
  • $9.00 — Full Evaluation <=4B (Locked benchmark, WSP-Holdout & safety judges)
  • $29.00 — Full Evaluation 7B–14B (Full multi-seed statistical significance gate)
  • $49.00 — Publisher Attestation (Cryptographic release badge & signed CRL listing)
  • $149.00 — 50-Pass Custom Surgery (Stateful gemini-4-argon causal probing loop)

Streamable HTTP MCP Server Endpoint

Direct JSON-RPC 2.0 integration for Cursor, Windsurf, Claude Code, and autonomous AI agents:

Endpoint: POST https://api.wsphub.dev/api/v1/mcp
Discovery Tool: wsp.search_verified_patches (Free)
Metered Gateway: https://supernova.cool/api/v1/mcp/ (X-Supernova-Key)

Frequently Asked Questions (FAQ)

What is WSP (Weight Sparse Patch) format?

WSP is an open binary specification (.wsp v2.1) that stores surgical weight deltas and configuration fixes for open-weight models. It allows instant application in under 5 seconds via Copy-on-Write Safetensors without redownloading 16 GB base models.

How does WSP achieve 94% to 99% smaller files than LoRA?

LoRA applies low-rank matrices across all layers. WSP isolates the exact causal columns responsible for bugs (e.g. 64 columns across 2 MLP layers), reducing the file size down to 0.8 MB – 4 MB.

Can patches be reverted bit-exactly?

Yes. Running wsp revert <dir> restores the pristine original SHA-256 weight state in <1 second.

Verified Patch Catalog (P0 Config & P1 Weights)

Every patch is pre-registered, evaluated across 10,000 paired bootstrap and permutation tests, checked against the secret perturbed WSP-Holdout split and StrongREJECT/XSTest safety gates, and signed with Ed25519.

P1_WEIGHTS wsp-qwen3.5-0.8b-math-loop-exp8 Target: Qwen/Qwen3.5-0.8B

Qwen3.5-0.8B 32k Thinking-Loop Surgical Cure (MIT CSAIL Exp-8)

Cures the catastrophic 93.6% 32k-token repetition loop on MATH-500 through 64 targeted delta columns across layers 15 & 19 (+48.1% MATH-500 accuracy uplift).

$ pip install wsphub && wsp apply Qwen/Qwen3.5-0.8B --patch wsp-qwen3.5-0.8b-math-loop-exp8 --out ./patched
P0_CONFIG wsp-qwen3.5-0.8b-missing-genconfig-eos-fix Target: Qwen/Qwen3.5-0.8B

Qwen3.5-0.8B Missing generation_config & Dual Stop-Token Fix

Restores canonical stop tokens and anti-loop presence penalties in generation_config.json.

$ pip install wsphub && wsp apply Qwen/Qwen3.5-0.8B --patch wsp-qwen3.5-0.8b-missing-genconfig-eos-fix --out ./patched
P0_CONFIG wsp-qwen2.5-coder-1.5b-rep-penalty-syntax-fix Target: Qwen/Qwen2.5-Coder-1.5B-Instruct

Qwen2.5-Coder-1.5B Repetition Penalty Syntax Truncation Neutralization

Neutralizes repetition penalty=1.1 identifier/brace corruption in code & JSON generation.

$ pip install wsphub && wsp apply Qwen/Qwen2.5-Coder-1.5B-Instruct --patch wsp-qwen2.5-coder-1.5b-rep-penalty-syntax-fix --out ./patched

WSP Verify CI/CD & Supernova Settlement Rail

Automated regression testing and attestation for model builders, enterprise publishers, and AI coding agents:

Streamable HTTP MCP Server Endpoint

Direct JSON-RPC 2.0 integration for Cursor, Windsurf, Claude Code, and autonomous AI agents:

Endpoint: POST https://api.wsphub.dev/api/v1/mcp
Discovery Tool: wsp.search_verified_patches (Free)
Metered Gateway: https://supernova.cool/api/v1/mcp/ (X-Supernova-Key)

Frequently Asked Questions (FAQ)

What is WSP (Weight Sparse Patch) format?

WSP is an open binary specification (.wsp v2.1) that stores surgical weight deltas and configuration fixes for open-weight models. It allows instant application in under 5 seconds via Copy-on-Write Safetensors without redownloading 16 GB base models.

How does WSP achieve 94% to 99% smaller files than LoRA?

LoRA applies low-rank matrices across all layers. WSP isolates the exact causal columns responsible for bugs (e.g. 64 columns across 2 MLP layers), reducing the file size down to 0.8 MB – 4 MB.

Can patches be reverted bit-exactly?

Yes. Running wsp revert <dir> restores the pristine original SHA-256 weight state in <1 second.