Releases·Tool Release
Turning GLM-5.3-Flash into a Jev-like decision model
A privatemode.ai blog post describes a prompting approach that gives standard LLMs like GLM-5.3-Flash Jev-like decision properties: crafting the input prompt so the first output token answers the question, enabling decisions in a single forward pass. The post details implementation with GLM-5.3-Flash and vLLM and benchmarks it against Jev and Laya.
172 engagement1 sourceSat, Sep 26, 2026, 03:49 PM
AI Overview
- privatemode.ai says prompting GLM-5.3-Flash to answer in the first output token yields Jev-like decisions in a single forward pass, benchmarked via vLLM against Jev and Laya
- The GLM-5.3-Flash setup matches Jev on accuracy and speed and substantially outperforms Laya, but Jev remains several times cheaper per decision
- Unlike Jev, the GLM-5.3-Flash setup supports vision inputs
Related
Releases·Tool Release
LLM2Jev turns local LLMs into Jev-style structured decision engines with prefill-only binary inference
A new open-source project, LLM2Jev, adapts local Hugging Face-compatible causal language models into Jev-style structured decision engines, emitting typed Choice, Score, and Noul answers with probabilities. It uses prefill-only binary inference — scoring candidates with yes/no logit judgments without generating answer tokens — and leverages SGLang's Radix Cache for shared-prefix efficiency. The project is independent and unaffiliated with Jev or TypeSafe.
A new open-source project, LLM2Jev, adapts local Hugging Face-compatible causal language models into Jev-style structured decision engines, emitting typed Choice, Score, and Noul answers with probabilities. It uses prefill-only binary inference — scoring candidates with yes/no logit judgments without generating answer tokens — and leverages SGLang's Radix Cache for shared-prefix efficiency. The project is independent and unaffiliated with Jev or TypeSafe.
109 engagement1 sourcegithubSat, Sep 19, 2026, 10:59 AM
Products·launched
jevos: open-source text classifier answering yes/no questions with a single forward pass
jevos is an open-source alternative to Jev: you send a text and a yes/no question and get an answer in one forward pass without generating text. It runs on a laptop CPU, making cheap yes/no classification over documents feasible. Aimed at developers who want fast, local rule-checking instead of full LLM generation.
jevos is an open-source alternative to Jev: you send a text and a yes/no question and get an answer in one forward pass without generating text. It runs on a laptop CPU, making cheap yes/no classification over documents feasible. Aimed at developers who want fast, local rule-checking instead of full LLM generation.
9 engagement1 sourceredditSat, Sep 26, 2026, 02:44 PM
Releases·Tool Release
Nokia Applied Research releases AnyJev, turning any LLM into a Jev-style typed-decision model with real probabilities and no training
Nokia Applied Research, with a Tencent Hunyuan collaborator, released AnyJev, an open-source Python package (Apache-2.0, on PyPI) that converts any LLM into a Jev-style decision model producing typed decisions and real probabilities with no training. The release demonstrates that raw logit readout is order-sensitive — Qwen3-8B flips a BANKING77 answer with 1.00 confidence when options are reversed — while AnyJev's L0 mode stays consistent with zero labels.
Nokia Applied Research, with a Tencent Hunyuan collaborator, released AnyJev, an open-source Python package (Apache-2.0, on PyPI) that converts any LLM into a Jev-style decision model producing typed decisions and real probabilities with no training. The release demonstrates that raw logit readout is order-sensitive — Qwen3-8B flips a BANKING77 answer with 1.00 confidence when options are reversed — while AnyJev's L0 mode stays consistent with zero labels.
144 engagement1 sourcegithubMon, Sep 21, 2026, 05:55 AM
Products·building
riderless: a decision API that reads logits instead of generating tokens
riderless is a FastAPI decision service that classifies questions into lettered options without generating any tokens. It runs prefill only on a Gemma 4 26B-A4B model via llama.cpp and returns the softmax distribution over A-Z label logits. This gives developers cheap, fast, probability-carrying decision calls from a stock instruction-tuned model on a single GPU.
riderless is a FastAPI decision service that classifies questions into lettered options without generating any tokens. It runs prefill only on a Gemma 4 26B-A4B model via llama.cpp and returns the softmax distribution over A-Z label logits. This gives developers cheap, fast, probability-carrying decision calls from a stock instruction-tuned model on a single GPU.
9 engagement1 sourceredditTue, Sep 22, 2026, 05:51 AM
Products·building
Gevva0: cascading LLM decision engine hitting ~2ms inference with GTO parity on blackjack
Gevva0 is a cascading decision engine that routes queries through tiers of increasing cost: static rules, absorbing Markov chains over embeddings, then a fast-path LLM forward pass. It uses EmbeddingGemma-2 for 768D query embedding and Gemma 4 MoE for the expensive tier. The goal is near-instant, accurate decisions (demonstrated on Vegas blackjack GTO parity) without paying full LLM latency for every query.
Gevva0 is a cascading decision engine that routes queries through tiers of increasing cost: static rules, absorbing Markov chains over embeddings, then a fast-path LLM forward pass. It uses EmbeddingGemma-2 for 768D query embedding and Gemma 4 MoE for the expensive tier. The goal is near-instant, accurate decisions (demonstrated on Vegas blackjack GTO parity) without paying full LLM latency for every query.
24 engagement1 sourceredditWed, Oct 14, 2026, 04:55 PM