Open inquiry. Rigorous methods.

Quark — A Conversational Assistant Built on Origin and NextGen

Technical report on Quark — system design, training, routing, evaluation, and responsible deployment of IdonAI’s primary assistant product.

2026-03-12 →

NextGen — Technical Report on Architecture, Training, and Evaluation

Full technical report on NextGen and Origin — architecture, training stack, Gemini-competitive benchmarks, ablations, scaling, and deployment considerations.

2026-02-18 →

Calibration and Abstention — Saying “I Don’t Know” Reliably

Probes and training signals for abstention when evidence is missing—progress and gaps on Origin and NextGen.

2026-01-20 →

2025

Safety Evaluation Card — 2025 H2 Snapshot

Compact safety scorecard for Origin, NextGen, and Quark—companion to the full Safety Research report.

2025-12-03 →

Agentic Tool Loops — Planning, Verification, and Stop Conditions

Experimental agent patterns on NextGen: plan–act–verify loops, and why Quark keeps tools conservative in production.

2025-11-14 →

Safety Research — Red-Teaming, Alignment, and Deployment Guardrails

IdonAI’s safety framework for Origin, NextGen, and Quark — threat modeling, red-teaming, evaluations, product guardrails, and limitations.

2025-10-22 →

Enterprise Isolation Patterns for Model Serving

Tenant isolation, VPC patterns, and audit logging considerations for IdonAI enterprise deployments.

2025-09-25 →

Over-Refusal Reduction without Safety Regressions

How we lowered benign over-refusal toward single digits while holding jailbreak metrics steady.

2025-08-07 →

Distilling NextGen into Origin — Quality without Latency Tax

Task-selective distillation that lifts Origin’s coding and MMLU scores while preserving Flash-tier serving.

2025-06-12 →

Prompt Injection Defenses in Retrieved and Pasted Content

Indirect prompt injection threat model, evaluations, and mitigations for Quark’s document-aware workflows.

2025-04-28 →

Structured Output Constraints — JSON Reliability for Tools

Constrained decoding, schema adherence metrics, and failure taxonomies for Origin and NextGen tool use.

2025-03-19 →

Benchmarks

Eval Harness Standardization — Templates, Seeds, and Leaderboard Hygiene

How IdonAI locks prompt templates and harness versions so Origin/NextGen numbers stay comparable across cuts.

2025-01-08 →

2024

Quark Routing Prototype — Dual-Model Assistant Control Plane

First Quark router design: features, policies, cascade mode, and cost–quality tradeoffs before public launch.

2024-11-21 →

Multimodal Alignment Pilot — Vision and Audio into the LM Space

Early encoder–LM alignment experiments that unlocked NextGen multimodal features.

2024-09-09 →

Continuous Batching Notes — Fairness, Prefill, and Cache Hits

Practical serving research: scheduling fairness between short chat and long prefill jobs on shared Origin pools.

2024-08-29 →

Mixture-of-Experts Routing Study for NextGen-Class Models

Expert count, top-k routing, and load-balancing ablations that shaped NextGen’s sparse FFN design.

2024-07-18 →

Long-Context Mid-Training — From 4K to 128K and Toward 1M

Curriculum and position-extension methods that enabled Origin 128K and paved the way for NextGen 1M.

2024-05-02 →

Datasets

Dataset Mixture Ablations — Code, Math, and Prose Tradeoffs

Systematic mixture ablations showing how code and math density reshape Origin-class downstream profiles.

2024-04-16 →

Red-Team Pack v1 — A Minimal Adversarial Suite for Release Gates

First standardized jailbreak and misuse suite used as a hard gate before Origin promotions.

2024-03-11 →

Preference Optimization v1 — Reward Models without Reward Hacking

First production preference-optimization stack: rater guidelines, KL control, and anti-sycophancy checks.

2024-02-20 →

2023

Origin Serving Canaries — Latency, Batching, and Failure Budgets

Operational research on validating Origin under load before Quark promotion: SLOs, canaries, and rollback.

2023-12-05 →

Origin — Technical Report on IdonAI's Fast Production Model

Full technical report on Origin: architecture, training pipeline, Flash-class benchmark suite, serving infrastructure, distillation methodology, safety interfaces, ablation study, and known limitations.

2023-11-16 →

Instruction Tuning Pilot — From Base LM to Helpful Assistant

First SFT mixtures and evaluation rubrics that turned early checkpoints into usable assistants—precursor to Quark behavior.

2023-07-09 →

Grouped-Query Attention for Interactive Serving

Why Origin adopted GQA: KV-cache memory, batchability, and quality tradeoffs measured on internal canaries.

2023-05-18 →

Tokenizer Design Notes for a Shared Origin–NextGen Vocabulary

How we chose a byte-level BPE vocabulary for code-heavy workloads and future multimodal text interfaces—fertility measurements, code round-trip tests, and the case for boring infrastructure done right.

2023-03-22 →

Datasets

Data Quality Filtering v1 — Provenance, Dedup, and Toxicity Screens

First production-oriented data pipeline for IdonAI pretraining: filter architecture, near-dedup strategy, toxicity screens, PII heuristics, and measured impact on training quality versus unfiltered baselines.

2023-02-14 →

2022

Early Scaling Observations on Small Dense Transformers

2022 internal study of loss-vs-compute on dense decoder-only models—what scaled predictably, what failed to transfer to Origin, and what design decisions these experiments permanently shaped.

2022-09-08 →