Curated, runnable demonstrations of the EML tool layer and the multi-agent investigations built on top of it. Every example in this directory is self-contained and runnable with python examples/<name>.py after pip install -e ..
| File | What it shows |
|---|---|
01_eml_basics.py |
The seven EML tools in one pass — translate, evaluate, verify, search |
02_hard_problems_via_sympy.py |
sympy_compute solving the five graduate-level hard problems |
03_riemann_zeros_on_critical_line.py |
Computes the first 20 non-trivial zeros of ζ and verifies they sit on Re(s)=½ |
04_riemann_li_criterion.py |
Computes Li's λₙ for n=1..10 and confirms positivity (RH-equivalent, Li 1997) |
05_riemann_robin_criterion.py |
Checks σ(n) < e^γ·n·log log n for 5040 < n ≤ 10000 (RH-equivalent, Robin 1984) |
06_riemann_pair_correlation.py |
Empirical pair correlation of normalized zero spacings vs Montgomery's GUE form |
pip install -e .
python examples/01_eml_basics.py
python examples/03_riemann_zeros_on_critical_line.py # takes ~1 min; high precisionAll six run offline and require no Claude API key. The subagent-orchestration flow (which needs Claude Code) is documented in ../RESULTS.md and ../RIEMANN_REPORT.md; the examples here are the pure computation half of those runs and can be used to verify the agents' numerical claims independently.
The exact prompts used to spawn Claude Code subagents for the benchmark and Riemann runs are embedded in ../src/eml_research/benchmark.py (for the general math benchmark via the Anthropic SDK) and in ../src/eml_research/riemann/README.md (for the RH investigation). See those files for the templates and the track-split (EML-first vs Classical-only) rules.