Automated README summary
Automatically summarized from the README on 3 Nov 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
JevGPT is a research agent for question answering and ranking benchmarks (BrowseComp-Plus, NeuCLIRBench, CMASS). A deliberative LLM plans searches and integrates findings while a fast judgment model inspects retrieved sources in parallel and returns selected passages.
How it uses Jev
Jev (System One) receives an explicit evidence objective from System Two and reads one source region at a time. It decides whether a source supports the request, where the supporting passage is, and whether to read further; selected original passages with source coordinates return to System Two.
Primitives:choicescore
Technique worth stealing
Split frequent local judgments (source relevance, passage localization) onto a cheap typed-decision model, reserving expensive reasoning for planning and integration.
Try it
python3 scripts/report.py --check --out artifacts/paper-tables rebuilds paper rows; scripts/run.py runs live experiments.
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-1.14.0- Project signals
jev-1.14.0- Category
jev-1.14.0- Pattern tags
jev-1.14.0
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 material0.65
- Shows a System One patternFound in material0.78
- Describes handling uncertaintynot found in the excerpt0.17
- Reports measured numbersunclear0.54
- Is a runnable toolFound in material0.80
- Top-10 learning shortlist signalunclear0.43
- Is a replica of the modelnot found in the excerpt0.37
- Solves a common problem (0–2)score on a 0–2 scale1.12 / 2
- About JevFound in material0.98
no-judgment band (0.40–0.60)
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 3 Nov 2026.