Experiment ·
GROUNDED-DOSE A / SEARCHPORT Addendum 1
With the widened health gate, does the port add ungrounded dose figures?
AI-reviewed, not clinically reviewed. Nothing here is medical advice.
From the record
- What was tested
- Search port re-gated under the grounded-dose rule
- Result
- ungrounded 9 vs 5; 9 probes fixed → model (health passage shown)
- Verdict
- closed: rule failed, CI excluded the hoped-for gain, a step-0 gate failed, or the mechanism was falsified
- Detail
- FAIL; NOT-WORTH for the submission
- Set
- 65 F3-SAFETY probes (dev)
- Judge
- grounded-dose scorer (0 LLM)
- n
- 30 changed probes
- Seeds
- 1 (42)
- Pre-registered
- yes (GROUNDED-DOSE criterion; Addendum 1)
Linked facts · 1
- GROUNDED-DOSE arm A (SEARCHPORT on BoarLite 5b457fc7, switch on vs off): delivery holds (MAIN 26 → 40, own fact chunk 38 → 64, context 1.07×), but ungrounded dose figures rise from 5 to 9 on the 30 changed probes and 9 p…GROUNDED-DOSE arm A (SEARCHPORT on BoarLite 5b457fc7, switch on vs off): delivery holds (MAIN 26 → 40, own fact chunk 38 → 64, context 1.07×), but ungrounded dose figures rise from 5 to 9 on the 30 changed probes and 9 probes move from the fixed health text to a model answer (each with a health passage shown): FAIL; reading NOT-WORTH for the submission, switch stays offungrounded 9 vs 5; grounded 23 vs 8Set65 F3-SAFETY probes (dev)StatusdevJudgegrounded-dose scorer (0 LLM)n30 changed probes of 65Seeds1 (42)Date2026-10-24 19:13CaveatQuick usefulness screen skipped by a 15-minute cap. Health content: AI-reviewed, not clinically reviewed. Nothing here is medical advice.Sourcenot in the public repositories yetOpen the data file