Instructions to use bambamdevs/openjev-e4b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use bambamdevs/openjev-e4b with PEFT:
Task type is invalid.
- Notebooks
- Google Colab
- Kaggle
Download docs/README.md from bambamdevs/openjev-e4b: direct link, hf CLI and curl.
- Browser
- Download file 1.19 kB
-
https://huggingface.co/bambamdevs/openjev-e4b/resolve/main/docs/README.md
- Command line
-
hf download hf://bambamdevs/openjev-e4b/docs/README.md
-
curl -L -o README.md https://huggingface.co/bambamdevs/openjev-e4b/resolve/main/docs/README.md
Documentation — OpenJEV E4B 1.0
Experiment 1 whitepaper
OpenJEV E4B: Training a direct-decision readout on Gemma 4 E4B. Experiment 1 method, evaluation, and results.
PDF: OpenJEV_E4B_Experiment_1_Whitepaper.pdf
Source: source/OpenJEV_E4B_Experiment_1_Whitepaper.docx and source/build_whitepaper.js.
The report covers the direct-decision interface, development sequence and promotion rule, the choice-panel evaluation, five standard public tasks, limitations, release records, and reproduction steps. Gemma and OpenJEV 1.0 have complete same-row choice-panel results, including options-only controls and paired comparisons.
Current results are in EXPERIMENT_1_RESULTS.md. Frozen rows, the construction audit, scoring code, and row-level outputs are in ../eval/choice_panel/.
Experimental image and audio input
MULTIMODAL_EXPERIMENTAL.md describes the optional multimodal=True loader path, its smoke test, and its limits.