jevbooks

← All projects

jev_robot

Hu-xiao-max/jev_robot

Local decider-2b picks robot skills via Jev-style typed choices.

Agent Decisions1.00Runner-up: Classification & Routing
Automated README summary

Automatically summarized from the README on 27 Sept 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.

What it is

A MuJoCo simulation where a local decision model controls an AgileX PiPER arm to push a red cube to three random positions and place it in a tray, recovering from interruptions.

How it uses Jev

Jev-style typed-choice interface: the model receives English sensor facts and returns one of ten fixed skills with confidence and full probability distribution. Deterministic checks can reject a choice but never substitute another; rejection triggers a new decision.

Primitives:choice

Technique worth stealing

Forward pass over decision labels instead of text generation, with deterministic execution checks that reject but never substitute.

Try it

Run .llm-venv/bin/python scripts/serve_decider.py --model models/decider-2b, then .venv/bin/python -m piper_sim --policy decider --moves 3 --output outputs/dynamic

View on GitHub

Start using Jev: API key and your first request →

Model judgments from README excerpts

Project signals use repository metadata, an existing summary and the first 700 README characters, not the full project. “Not found” means not detected in that material; it does not establish absence. These are the original judgments with recorded models and dates, not source audits or run results.

Models & evaluation dates

Inclusion gate
jev-1.13.0
Project signals
jev-1.13.0
Category
jev-1.13.0
Pattern tags
jev-1.13.0

Evidence

These signals use repository metadata, an existing summary and at most 700 cleaned README characters. ≥ 0.60 means reported in that excerpt; ≤ 0.40 means not found there. The middle band is inconclusive. None is a source-code audit or a test result. The runnable-tool question excludes research experiments; the recommendation question asks for a learning shortlist, not a quality verdict.

  • Jev is core to the projectFound in material0.92
  • Shows a System One patternFound in material0.75
  • Describes handling uncertaintynot found in the excerpt0.12
  • Reports measured numbersnot found in the excerpt0.05
  • Is a runnable toolunclear0.44
  • Top-10 learning shortlist signalunclear0.47
  • Is a replica of the modelFound in material0.61
  • Solves a common problem (0–2)score on a 0–2 scale0.16 / 2
  • About JevFound in material0.90
Signals
Jev is core to the project0.92Shows a System One pattern0.75Describes handling uncertainty0.12Reports measured numbers0.05Is a runnable tool0.44Top-10 learning shortlist signal0.47Is a replica of the model0.6100.51Solves a common problem (0–2)0.16 / 2012

no-judgment band (0.40–0.60)

Design patterns
Inspection0.30Funnel0.28Check-up0.2000.51

no-judgment band (0.40–0.60)

The Jev model and date for each judgment stage are listed above. Card written by DeepSeek V4.1 Flash from the README on 27 Sept 2026.