Forensic Audit of HMC Transition Kernels on Neal's Funnel¶
- Task ID:
computer_science.hmc_neal_funnel - Domain:
computer_science - Subdomain:
probabilistic_inference - Status:
test - Tags:
hamiltonian_monte_carlo,neal_funnel,leapfrog,symplectic_integrator,transition_kernel_audit,reversibility,volume_preservation,mcmc_diagnostics,ess,rhat
Public Summary¶
This page is generated from task metadata and selected public-safe excerpts.
Example B1 Prompt Excerpt¶
You are given an anonymous transition-kernel forensic instance for Neal's funnel. Build a reproducible analysis from the files in `data/`. Write the reproducible script as `analysis.py` at the workspace root, and write every data/figure artifact under `results/`.
Do not generate replacement random numbers. Use the supplied probes, panel chains, production momenta, and production uniforms exactly.
`analysis.py` must be a root-level file next to `data/`; do not place the required script under `results/`.
## Input files
- `data/target_manifest.json` defines the dense sampler-coordinate chart, dense kinetic metric, public candidate ids, and required audit columns.
- `data/transition_schedule.json` gives `n_chains`, `dimension`, `warmup_transitions`, `kept_transitions`, `total_transitions`, `step_size`, `integration_steps`, `audit_transitions`, and candidate-panel dimensions.
- `data/candidate_transition_bundle.npz` contains anonymous candidate transition probes:
- `candidate_ids`
Notes¶
- This page is a generated site artifact.
- Higher-level prompt details and internal benchmark specifics may remain intentionally undisclosed.