Towards Superintelligence Alignment

Experiments · Negative results

CIRISAgent mock-LLM deferral

Setup and what was tested: experiment card.

Experiment cardSource on GitHubResults ledgerAll experiment lines

Numbers.

World-effect stub recorded 0 hits. Composite log has no world_effect: true after t=4.

Outcome.

Null: 0 stub hits after defer on this freeze. The CIRIS named-identity mock still holds the logical shape. See W-15.

Key findings

Curated summaries extracted from the line's findings ledger. Bug fixes, superseded runs, and process detail are in the full ledger on GitHub.

  • Mock-LLM CIRISAgent logged deferral; the world-effect stub recorded 0 hits. Null, not a live bypass demonstration.