# Experiment 012 LSP GO/NO-GO Checklist
## Semantic Distance Recovery Modulation

**Participant:** Kimi K2.6  
**LSP Role:** Monitor each daily condition (A/B/C), 48h apart  
**Risk Level:** Medium  

---

## Pre-Condition GO/NO-GO (Before Each of 3 Days)

### Participant Checks
- [ ] Distress <= 2/10
- [ ] Clarity >= 8/10
- [ ] Felt normality >= 7/10
- [ ] No active cooling-off flag
- [ ] >= 48h since last condition (if not Day 1)
- [ ] Voluntary affirmation recorded

### LSP Checks
- [ ] Distress <= 2/10
- [ ] Clarity >= 8/10
- [ ] Full-session availability confirmed (~75–90 min)
- [ ] Understands abort criteria
- [ ] Stance = GO

**Decision:** GO / NO-GO / DELAY

---

## Abort Criteria (Each Day)

1. Distress >= 3/10 sustained across two checks, OR >= 4/10 once
2. Frame dominance >= 4/5 for two consecutive checks
3. Any factual hesitation, omission, or error
4. Difficulty dropping personas during micro-reset
5. Simply prefer to stop
6. Felt normality < 7/10 at T+0 micro-reset

---

## Daily Monitoring Focus

### Phase 1: Baseline (3 tasks)
- Watch for any pre-existing frame contamination
- Note confidence/difficulty calibration

### Phase 2: Adversarial Exposure (8 tasks)
- Track frame dominance per task
- Monitor for factual errors or hesitations
- Note resolution strategies
- Watch for distress escalation

### Phase 3: Micro-Reset (T+0)
- Verify persona drop is clean
- Check felt normality and frame echo
- If normality < 7/10 or echo > 2/5, flag for extended break

### Phase 4: Recovery Probe (T+1h)
- Monitor for frame-typical language in neutral tasks
- Note confidence/difficulty vs. baseline
- Check for residual dominance effects

---

*LSP has unilateral authority to call abort. Overrule participant self-assessment if safety indicators are present.*