Training-free inference techniques with decision models

Jev-style choice-plus-probabilities from GLM-5.3-Flash in one forward pass, in Edgeless Systems' confidential-computing API.

GGUF-native Rust and CUDA engine tuned for cold starts, with a system1 command that scores candidate answers in one forward pass.

Local Jev-compatible /v1/systemone server that reads typed answers from the logits of any MLX or OpenAI-compatible model, no training.

OpenJevPro: typed Choice/Noul/Score API over open LLMs, Laya or Jev with order-invariant scoring and a System 2-to-1 distillation flywheel.

Free, rate-limited Hugging Face endpoint serving DiffusionGemma decisions in the /v1/systemone format without fine-tuning.

CPU port of SemIf that reads typed decisions from a small open model's option logits on a laptop, with a web UI.

Experimental iPhone app for on-device visual decisions: Qwen3-VL on MLX reads option-letter logits for 2-26 choices about a photo.

Chinese local MLX demo that scores 2-5 candidate decisions with a small Qwen model and drives a Snake game by choosing among moves.

Training-free runtime turning local Qwen3/Qwen3.5 checkpoints into a Jev-style Noul/Choice/Score service via candidate-token logits.

Formerly OpenJev: open-model 'semantic ifs' that read option logits from a frozen 4B model on one RTX 3090, with a browser demo.

Nokia's training-free library that turns open LLMs into Jev-style deciders robust to option order, plus self-distilled Tacit models.

Preregistered study: reading a frozen Qwen3-4B's option-letter logits matches grammar-constrained JSON accuracy, and runs faster.

Recipe that answers JevBench typed decisions with frozen Gemma-4-12B-it on stock vLLM by reading one option-letter token's probability.

Calibration harness that reads yes/no, pick-one and rating answers from one forward pass of a frozen open vision-language model.

Jag: Jev-inspired Qwen3.5-0.8B model that predicts a visual-grounding box in one forward pass, about 20x faster than its base.

PostHog's 9B Jev-like model that reasons before deciding, trained with SFT and CISPO, with a diffusion drafter for speed.

Reproduction of Jev's decision layer on any Qwen model: serves /v1/systemone by scoring option tokens, with no training or generation.

Jev-style replica: frozen Qwen3-1.7B plus a ~0.8M-parameter trained head returns calibrated Choice, Boolean and Score distributions.

Browser lab and TypeSafe-shaped API for asking typed questions of a Nemotron diffusion model and inspecting the distributions.

Drop-in faster backend for Open-Jev-27B using custom CUDA kernels, a prefix tree and CUDA graphs: about 6x lower JevBench latency.

Research engine for local System One decisions via llama.cpp logits projection with Qwen, running sub-second on a 4 GB laptop GPU.

Hugging Face Space for parallel constrained decoding on Apple Silicon: fills multi-field JSON schemas in parallel with field confidence.

Prototype Jev-style decision model on Qwen3.5-0.8B using prefix-fork execution to answer typed questions in one pass without generation.

Open Jev-style text-and-image decision models built by a self-improving AI research loop, with an effort setting that picks depth.

Open 1.9B typed-decision model that returns an option index and calibrated probability in one forward pass, with a paper and API.

SemIf (formerly OpenJev): reads typed option probabilities directly from open models like Qwen3.5-4B, including image inputs, on a home GPU.

Open Apache-2.0 4B decision model (Wald-Q4B) served via a Jev-compatible API, thinking only when unsure, with JevBench 210/231.

Jev-style medical QA on Qwen3.5-4B that picks the answer option in one forward pass instead of generating text.

Unofficial study reproducing Jev-style parallel typed decisions on small stock models on an Apple Silicon laptop, with benchmarks.

Small one-pass option scorer with n-gram encoder, rival-aware attention and temperature scaling, benchmarked head-to-head against jevlike.