Calibration & Research

Laptop-scale reimplementation of RL for Calibrated Decisions: parallel constrained decoding trained with GRPO and a Brier reward.

Jev-class decision server on an RWKV-7 1B checkpoint that reads typed answers from a constrained softmax behind /v1/systemone.

Open Jev replica on Gemma 4 E2B and Gemma 3 270M served on Modal, with eight demos and a TypeSafe eval head-to-head.

Lightweight reproduction of the Jev shape on a ModernBERT encoder and a diffusion LLM, comparing speed, accuracy, calibration and cost.

Write-up with measurements on calibrating Jev as a code reviewer: context effects, textual tells, thresholds and noise near thresholds.

Local System One-compatible server that reads option probabilities from a small MLX LLM, with measured calibration.

Standalone Jev-class emotion model: one pass over a message sequence returns calibrated PAD scores, an emotion family and yes/no judgments.

Open project training a vision-enabled Jev-like model (0.8B, 9B) that judges dynamic candidate sets over images with Choice, Noul and Score.

Self-hostable calibrated decision model for images and text (Qwen3-VL-4B + LoRA) scoring 82.7 percent on public JevBench.

pijev: drop-in wrapper for the TypeSafe Python SDK that averages Jev answers across option orders in one request.

Calculator with no arithmetic in its code: Jev picks each answer character from 13 options, exposing its distributions and jitter.

Calibration experiments showing Jev's Choice probabilities on fair dice and coins stay high while accuracy stays at chance.

Learning project that builds choose-an-answer decision models from scratch and from Qwen, with downloadable 0.8B and 4B releases.

Ollama-style single binary that pulls and runs open decision models locally, with a TypeSafe-compatible API and playground.

Local, deterministic Rust decision engine reading single-letter answer logits from one llama.cpp pass, wire-compatible with Jev.

Playful web arena where Jev does arithmetic, writes chat replies and draws pixel art one Choice or Noul decision at a time.

Chat interface with no generative model: Jev picks intent, stance, claim and tone, and a deterministic compiler writes the reply.

Toy local Jev-shaped decision API on a 0.5B model that answers yes/no, choice and score questions by scoring option tokens.

Independent Jev-style server on DiffusionGemma and SGLang that classifies doodles, flowers and messages with Choice, Score and Noul.

Small experiment: reordering enum options changed a local Jev clone's answers on its old default, while hosted Jev's did not.

Python harness stress-testing Jev's confidence, its accuracy as a RAGTruth hallucination judge, and its use as a RAG reranker.

Toolkit for reading Choice, Score and Noul decisions out of embedding models, with data, fine-tuning, benchmarks and a server.

Audio-native Jev-style model: a frozen Whisper encoder encodes speech once and typed questions return emotion and prosody distributions.

Experimental Jev wire-compatible engine that claims to produce typed decisions from Mandelbrot fractal dynamics rather than a trained model.

Cloudflare's Apache-2.0 27B multimodal decision model that scores every option of every question jointly, with a Jev-compatible API.

Fastino's 340M Apache-2.0 decision classifier that scores caller-defined labels in one forward pass for routing, intent and scoring.

Perplexity's Apache-2.0 multimodal decision model on Qwen3.8-27B, reporting 61.56 on the Decision Index versus 57.9 for Jev.

Multilingual non-autoregressive decision engine answering choice, score and yes/no questions in one ~33 ms pass, with a checkpoint router.

Family of small Jev-like decision models you can train and run yourself, with a server compatible with TypeSafe's SDK.

Open-source Jev-compatible decision model served by one native binary on a laptop CPU, returning probabilities in 28 to 130 ms.