All projects

jev-baselines-eval

ickma2311/jev-baselines-eval

Pre-registered independent eval of TypeSafe Jev against a nano-class LLM, a frontier LLM, and a supervised encoder (Banking77 + CLINC150 zero-shot)

Python
Stars
6
Forks
0

Why it is listed

Independent pre-registered studies of TypeSafe Jev intent classification and confidence-gated cascades. Listed for evaluation evidence and baseline methodology, not validation of every published result.

Related projects

  • typed-decision-bench

    Bench of typed decision models: Jev vs OpenJev vs Laya, small local LLMs and cheap hosted LLMs on the same zero-shot classification tasks, in English and French.

    Shared topics: llm-evaluation

  • jev-2048

    An instrumented 2048 web lab where every move is a Jev (TypeSafe AI System One) Choice, with no heuristic fallback | 用 Jev 决策模型驱动每一步的 2048 网页实验台,概率、置信度、延迟与成本全部摊开可见,且刻意不做启发式兜底

    Shared topics: llm-evaluation

  • jev-testbench

    A lightweight, auditable Jev testbench built through Agentic Engineering, with frozen evidence, a preregistered null result, and a bilingual technical blog. 基于 Agentic Engineering 的轻量、可审计 Jev 实验台,包含冻结证据、预注册负结果、双语技术博客与认知负债反思。

    Shared topics: preregistration