YenkLabs

// Current Focus

⚡ Dali — the open verification layer for AI: it creates, scores, and preserves evidence so AI-assisted outputs can be independently verified, exchanged, and replayed. Current work: Open Evidence Corpus, seed verification sample, five-outcome taxonomy, Evidence Package Specification (EPS) v0.1, and the full evaluation run on data/results/ (early-stage v0.2.1).

// Dali Evidence Ecosystem

                    Dali
        Evidence Infrastructure Platform
                     │
─────────────────────────────────────────
Evidence Corpus · Benchmarks · Taxonomy
Evidence Packages · Replay Engine · APIs

Public assets today — corpus, seed sample, taxonomy, full evaluation run — connected by a single idea: evidence, not failure. Failures are seed data. The seed sample is for methodology review. The full run is the measured result set. Dali is the engine that verifies and preserves it.

Full ecosystem index: huggingface.co/yenklabs · Engine: github.com/yenklabs/Dali

// Recent Activity

Jul 11 Full evaluation run: 524 citations, 3 models, 5 jurisdictions — github.com/yenklabs/Dali/data/results
Jun 29 Renamed HF dataset: open-evidence-corpus (Dataset Viewer live)
Jun 29 Draft: Evidence Package Specification (EPS) v0.1
Jun 8 Published: Mata v. Avianca reconstruction
Jun 8 Published: open-evidence-corpus, dali-citation-benchmark, dali-verification-taxonomy
Jun 8 Released: Citation Benchmark seed sample (5 cases / 14 authorities)
Jun 7 Added 5 seed failure records to open evidence corpus
Jun 6 Released: Dali verification taxonomy

// The Forge

dali [flagship] → Python

Dali is the open verification layer for AI: it creates, scores, and preserves evidence so AI-assisted outputs can be independently verified, exchanged, and replayed. Early-stage v0.2.1.

yenk-inference-bench → Python

planned — Benchmarking local and hosted language models across latency, throughput, context utilization, and verification-oriented workloads.

micro-transformer-scratch → C++ / CUDA

planned — A transformer built from first principles to explore attention, training dynamics, memory movement, and inference mechanics without framework abstractions.

// Public Assets

Full evaluation run — available · 524 citations, 3 models, 5 jurisdictions · LEADERBOARD.md
Evidence Package Specification (EPS) v0.1 — available · draft interchange format · OpenAPI for evidence
Open Evidence Corpus — available · 5 seed records · open-evidence-corpus
Citation Benchmark (Seed Sample) — available · 5 cases, 14 authorities · dali-citation-benchmark
Verification Taxonomy — available · 5 outcome classes · dali-verification-taxonomy
Mata Reconstruction — available · investigation
Park Reconstruction — available · evidence record
Ross Reconstruction — planned
Research models (taxonomy / citation risk / authority / proposition) — planned

Ecosystem mirrored on Hugging Face: yenklabs

// Platform Stack

Open Evidence Infrastructure

Public corpora, EPS, Hugging Face datasets, and open-source verification engine.

Evidence Corpus

Portable evidence records — web browser · open-evidence-corpus on Hugging Face.

Verification Engine

Taxonomy, benchmarks, replay determinism, cryptographic lineage — Dali.

Enterprise Platform

Commercial evidence layer for regulated workflows — forthcoming.