Answers are grounded in this paper and cite their sources.
The pre-registration ledger of Nous, a PhD experiment on an LLM multi-agent energy-community simulator (three communities with private supply contracts, an LLM aggregator, announced grid outages, import surcharges and advisory notices; a pluggable decision-maker; two local model families). Every prediction about a run was written down, with its date, before the run; this document renders the whole ledger as it stands — each prediction's id, registration date, claim, read and status — with the registered drops and the blocks still to run. Version 1, generated 2026-10-27: 44 registered predictions, 0 not yet run (none). Registry checksum at generation (sha256 of docs/thesis_cells.json): e4d70a0df416a38822b608b3384e7e24f3dbef6ea7b1d148525c67398ab1c5f6; git f41ac8d; file sha256 b856b0d11c890ab6e6b9df8caf1b3484b6f1d767adefefd3c1eaa56822602406. A version deposited before a block runs is evidence that the block's predictions preceded its results; a version after each gate stamps the reads. Claims and dates are never edited between versions — only reads and statuses are added. Generated by baselines/preregistration.py in the author's repository.
No takes yet. Share an insight, caveat, or question.
Enea Mele (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: