Research Provenance — Source Verification Log
This page tracks the verification date and status of every external source cited across the hwLedger research briefs. All sources were verified on April 19, 2026.
2026 Models & Architecture (April 2026)
| Source | Type | Status | Last Verified | Notes |
|---|---|---|---|---|
| Meta Llama 4 Multimodal Intelligence | Blog | ✅ Active | 2026-04-19 | Llama 4 Maverick: 17B active, 400B total, iRoPE, sparse MoE routing |
| Gemma 3 Technical Report | arXiv | ✅ Active | 2026-04-19 | 5:1 interleaved local (1024-token window) + global attention, 128K context |
| Mamba-3: Improved Sequence Modeling | arXiv | ✅ Active | 2026-04-19 | MIMO variant: state_size=64 (2× reduction vs Mamba-2), Mar 2026 |
| DeepSeek-V3 Model Documentation | HF Docs | ✅ Active | 2026-04-19 | MLA parameters: kv_lora_rank=512, qk_rope_head_dim=64 |
| Qwen 3.6 GitHub Repository | GitHub | ✅ Active | 2026-04-19 | Hybrid GDN + sparse MoE, 256 experts, 1M context |
| Jamba-1.5: Hybrid Transformer-Mamba at Scale | arXiv | ✅ Active | 2026-04-19 | 94B active (Large), 12B active (Mini), 256K context, hybrid layers |
Inference Engines (April 2026)
| Source | Type | Status | Last Verified | Version |
|---|---|---|---|---|
| vLLM v0.19.0 Release Notes | GitHub | ✅ Active | 2026-04-19 | v0.19.0 (Apr 2026): PagedAttention v2, MLA support, Hugging Face integration |
| oMLX Repository | GitHub | ✅ Active | 2026-04-19 | v0.3.6 (Apr 2026): SSD-paged KV cache, MLX wrapper, MLA support |
| Apple MLX Framework | GitHub | ✅ Active | 2026-04-19 | Actively maintained, MLA kernels, unified memory optimizations |
| mistral.rs Releases | GitHub | ✅ Active | 2026-04-19 | MLA support (native), CUDA+Metal, MoE-aware routing |
| llama.cpp | GitHub | ✅ Active | 2026-04-19 | Universal GGUF format, partial MLA support, ROCm/CUDA/Metal backends |
Attention Research Papers
| Source | Type | Status | Last Verified | Year | Notes |
|---|---|---|---|---|---|
| Llama 2: Open Foundation and Fine-Tuned Chat Models | arXiv | ✅ Active | 2026-04-19 | 2023 | Foundation for GQA adoption |
| Mistral 7B | arXiv | ✅ Active | 2026-04-19 | 2023 | Sliding window attention pioneer |
| Mixtral of Experts | arXiv | ✅ Active | 2026-04-19 | 2024 | Sparse MoE standard |
| GQA: Training Generalized Multi-Query Transformers | arXiv | ✅ Active | 2026-04-19 | 2023 | Grouped-query attention formalization |
| Mamba: Linear-Time Sequence Modeling with Selective State Spaces | arXiv | ✅ Active | 2026-04-19 | 2023 | SSM foundation |
| Efficient Streaming Language Models with Attention Sinks | arXiv | ✅ Active | 2026-04-19 | 2023 | Attention sink mechanism |
FFI & Platform Tools (April 2026)
| Source | Type | Status | Last Verified | Version |
|---|---|---|---|---|
| UniFFI v0.31.0 | GitHub | ✅ Active | 2026-04-19 | v0.31.0 (Jan 2026) |
| CXX-Qt v0.7 | GitHub | ✅ Active | 2026-04-19 | v0.7 (KDAB, 2026) |
| Slint 1.15.1 | GitHub | ✅ Active | 2026-04-19 | v1.15.1 (Apr 2026) |
Competitive Tools & References
| Source | Type | Status | Last Verified | Notes |
|---|---|---|---|---|
| HF Model Memory Usage Calculator | Web | ✅ Active | 2026-04-19 | Streamlit-based; no MLA/Mamba support |
| can-it-run-llm Space | Web | ✅ Active | 2026-04-19 | Community-maintained; GPU selector |
| LM Studio | Desktop | ✅ Active | 2026-04-19 | Cross-platform GUI; GGUF models |
| Claude Opus 4.7 Announcement | Blog | ✅ Active | 2026-04-19 | Released Apr 2026; 2576px (3.75MP) image support, extended thinking |
Verification Notes
- All links verified as resolvable via HTTP HEAD requests on 2026-04-19.
- arXiv papers confirmed via direct abstract pages.
- GitHub repositories confirmed as actively maintained (commits within last 30 days).
- Blog posts from official vendor sources (Meta, Anthropic, Google, etc.).
- Withdrawn or archived papers: None detected in citation set.
- Dead links: None. All sources remain accessible as of 2026-04-19.
Next Refresh Cycle
Expected: June 15, 2026 (8-week interval)
Tasks:
- Re-verify all URLs for accessibility.
- Check for new model releases (Llama 5, DeepSeek-V4, Gemma 4, Qwen 4.x).
- Update vLLM/mistral.rs/llama.cpp version pins.
- Audit new papers on arXiv for attention mechanisms, quantization, MoE scaling.
Generated: 2026-04-19 UTC
Last Verified: 2026-04-19 UTC
Brief Coverage: All 12 research briefs + KV-cache math page