spb/llmindex

Public

The discriminative, contamination-resistant, fully transparent LLM ranking — updated live.

webapp ai llm benchmark

8commits
1branches
0releases
1.4 MBon disk
maindefault branch
2 mo agolast push

Research Mode is not enabled for this repository — this is a preview built from detected signals (1 LaTeX, 1 PDF, 1 BibTeX, 0 figures). Enable it with PATCH /api/v1/repos/llmindex {"research": true}.

1manuscripts (PDF)
1LaTeX sources
0figures
0tables
1experiment outputs
0versions

Manuscript

Rendered by your browser from the repository's own PDF — nothing is compiled on the server.

docs/whitepaper/llmindex-whitepaper.pdf412.6 KB Download Open

LaTeX sources

Main documents are detected by \documentclass; metadata is parsed, never compiled.

-2.2em [0.9em] [0.35em] A Psychometric, Contamination-Resistant, and Fully Auditable Methodology for Ranking Large Language Models [0.6em] Methodology White Paper --- Index Version

docs/whitepaper/llmindex-whitepaper.tex · article

4 figure env.2 table env.35 citations bib: references

docs/whitepaper/llmindex-whitepaper.tex49.4 KB
docs/whitepaper/references.bib9.7 KB

Figures 0

No figures directory

Images under figures/, fig/, plots/ or img/ appear here.

Tables 0

No tables/ directory. Table environments inside LaTeX sources are counted above.

Reproducibility checklist

Static checks on the file tree at main. Nothing is executed.

yes README with instructions README.md
yes Declared environment (requirements / environment.yml / pyproject / lockfile) apps/psychometrics/requirements.txt
yes Pinned dependencies (lockfile) pnpm-lock.yaml
no Container definition (Dockerfile / Apptainer)
yes Pipeline entry point (Makefile / justfile / Snakefile / dvc.yaml) apps/psychometrics/Makefile
no Scripts directory
yes Data directory or data access notes 2 files
yes Automated tests 5 files
yes Explicit random seeds (file names) 1 files
no License
no Citation metadata (CITATION.cff / .bib)

Experiment outputs 1

data/runs/README.md320 B

Manuscript versions

Tags link the paper to the exact code and data snapshot.

No tagged versions yet

Tag each submitted version: git tag v1-submitted && git push --tags