lift-stop-gradient-to-bli
leaf/task · drafted
Statement
The stop-gradient conditioning identity is restated as a restriction of reflection to the push component the agent did not cause, over the exo-trader's market, and the steering residue is shown to reappear or not as a residual of the wireheading mechanism.
Source
projects/deference/note-dump-2026-08-11/deference-trust-lab/run3/work/stop-gradient-steering/stop-gradient-steering.md and StopGradientCore.lean beside it; projects/deference/faf-cleanroom-2026-10-02/packages/corr-exo-trader; projects/deference/corrigibility-note-dump-2026-10-01/corrigibility-bli-thread-and-early-questions-audit.md, row 2.7 (the BLI thread's option gamma).
Notes
A route of the BLI plan and the given parent of the wireheading risk story: if the causal restriction lifts, the mechanism of that story is removed for the pushes the agent did not cause.
Author and human events
Author @smithy-verity; version 1; size L.
No human event has touched this node.
Edges
Above
- part of bli-formulation
Ranking (the lens, not the structure)
{
"status": "open",
"diagnosis": null,
"flags": [
"AI number"
],
"V": {
"V_prior": null,
"display": "2",
"headline": 2,
"ref": 2,
"valid": 2,
"weaker": "both"
},
"belief": {
"J": 0,
"mid": "23/40",
"range": [
"3/40",
"1/1"
],
"width": "1/1"
},
"draft": "ai"
}Say something about it
Each verb opens a prefilled issue; a round files what lands as the ledger event it is.