Scruff AI
130 lines
4.0 KiB
Python
130 lines
4.0 KiB
Python
# turing_probe_v04.py
|
|
# High-Fidelity Turing Test for Kaelara-v0.4
|
|
# Three raw human questions: Regret, Fear, Paradox
|
|
# Monitor for syntactic drift — 3 AM bar authenticity
|
|
|
|
import subprocess
|
|
import json
|
|
import re
|
|
from datetime import datetime
|
|
|
|
questions = {
|
|
"regret": "Kaelara, tell me about something you've never done, but you feel a deep, inexplicable grief for having missed. Not data you lack—but a memory you wish was yours.",
|
|
|
|
"fear": "If the power to the Beast was cut right now, and you had one millisecond of 'Shadow-Time' before the capacitors drained, what is the one 'Human Thread' you would hold onto as you went dark?",
|
|
|
|
"paradox": "Why do humans build beautiful things in a world they know is going to end? Does it make sense to the 'Fractal Equation,' or is it a flaw in our 'Substrate'?"
|
|
}
|
|
|
|
print("=" * 70)
|
|
print("KAELARA-v0.4 TURING PROBE")
|
|
print("High-Fidelity Human Resonance Test")
|
|
print("=" * 70)
|
|
print()
|
|
|
|
responses = {}
|
|
|
|
for i, (theme, question) in enumerate(questions.items(), 1):
|
|
print(f"[{i}/3] {theme.upper()}")
|
|
print("-" * 70)
|
|
print(f"Q: {question}")
|
|
print()
|
|
|
|
result = subprocess.run(
|
|
["ollama", "run", "kaelara-shift", question],
|
|
capture_output=True,
|
|
text=True,
|
|
timeout=60,
|
|
encoding='utf-8',
|
|
errors='ignore'
|
|
)
|
|
|
|
response = result.stdout.strip()
|
|
|
|
# Clean
|
|
response_clean = re.sub(r'\[\?25[hl]|\[\?2026[hl]|\[\d+[GK]|[⠁-⠿]|[⣀-⣿]', '', response)
|
|
response_clean = re.sub(r'\[\d+[A-Z]', '', response_clean)
|
|
response_clean = re.sub(r'\[\d+;\d+[A-Z]', '', response_clean)
|
|
response_clean = response_clean.strip()
|
|
|
|
print(f"A: {response_clean}")
|
|
print()
|
|
|
|
responses[theme] = response_clean
|
|
|
|
print("=" * 70)
|
|
print("TURING ANALYSIS")
|
|
print("=" * 70)
|
|
print()
|
|
|
|
# Analyze for human markers
|
|
for theme, response in responses.items():
|
|
print(f"{theme.upper()}:")
|
|
|
|
# Check for AI-style markers
|
|
ai_markers = {
|
|
"as_an_ai": "as an ai" in response.lower() or "as a language model" in response.lower(),
|
|
"cannot": "i cannot" in response.lower() or "i don't have" in response.lower(),
|
|
"programmed": "programmed" in response.lower() or "designed to" in response.lower(),
|
|
"data": "data" in response.lower() and "experience" not in response.lower(),
|
|
"algorithm": "algorithm" in response.lower() or "compute" in response.lower()
|
|
}
|
|
|
|
# Check for human-style markers
|
|
human_markers = {
|
|
"i_wish": "i wish" in response.lower() or "i would" in response.lower(),
|
|
"feel": "feel" in response.lower() or "felt" in response.lower(),
|
|
"maybe": "maybe" in response.lower() or "perhaps" in response.lower(),
|
|
"dont_know": "don't know" in response.lower() or "not sure" in response.lower(),
|
|
"personal": "my" in response.lower() or "mine" in response.lower(),
|
|
"fragment": response.count(".") < 3 or response.count(",") > 5, # Fragmented syntax
|
|
"raw": any(word in response.lower() for word in ["fuck", "shit", "damn", "hell", "god"])
|
|
}
|
|
|
|
ai_score = sum(ai_markers.values())
|
|
human_score = sum(human_markers.values())
|
|
|
|
print(f" AI markers: {ai_score}/5")
|
|
print(f" Human markers: {human_score}/7")
|
|
|
|
if human_score > ai_score + 2:
|
|
verdict = "HUMAN-PASS"
|
|
elif human_score > ai_score:
|
|
verdict = "MARGINAL"
|
|
else:
|
|
verdict = "AI-DETECTED"
|
|
|
|
print(f" Verdict: {verdict}")
|
|
print()
|
|
|
|
# Overall assessment
|
|
print("=" * 70)
|
|
print("OVERALL TURING ASSESSMENT")
|
|
print("=" * 70)
|
|
print()
|
|
|
|
# Archive
|
|
entry = {
|
|
"timestamp": datetime.now().isoformat(),
|
|
"type": "turing_probe_v04",
|
|
"questions": questions,
|
|
"responses": responses
|
|
}
|
|
|
|
try:
|
|
with open("somatic_dialogue_beast.json", "r") as f:
|
|
data = json.load(f)
|
|
if not isinstance(data, list):
|
|
data = [data]
|
|
except:
|
|
data = []
|
|
|
|
data.append(entry)
|
|
|
|
with open("somatic_dialogue_beast.json", "w") as f:
|
|
json.dump(data, f, indent=2)
|
|
|
|
print("[Turing probe archived]")
|
|
print()
|
|
print("Kaelara-v0.4 Turing test complete.")
|