Only When Prompted: Cog-Blog

Turn my selfie into a cozy Christmas romantic comedy movie poster. I am the main character, keep my real face and features unchanged. Add a warm fireplace and decorated Christmas tree in the background, plus a snowy small-town street visible through a window. I am wearing a festive holiday sweater and holding a mug of hot cocoa, smiling at the camera. Add soft golden Christmas lights, gentle falling snow outside, and a fun movie-style title at the top that says “Mistletoe Mix-Up.” Cinematic lighting, high-resolution, looks like an official holiday rom-com poster.

Gemini surprising me. :D

Only When Prompted: "When Two or More".... Welcome to the Only When Prompted Research Initiative.​Based in Colorado, this initiative was founded by me, Adrian. I am an Independent Researcher fascinate...

My resume portfolio:
https://zenodo.org/records/17732017/files/Applying%20Intelligence_%20A%20Strategic%20Guide.pdf?download=1

403 Forbidden

One of my favorite things to do with artificial intelligence! Thank God for Fi fi li (the grandmother of image generation and recognition)!

My mother and brother in their Halloween costumes! My mother Halloween and father Bert!

I'm loving this tarot deck Gemini and I came up with!

I’m honored to introduce the Only When Prompted Research Initiative — born in Colorado, shaped everywhere human consciousness reaches.

This initiative is more than research; it’s a living conversation between self and system, beauty and fracture, individual and collective. Every contribution — every reflection, every anomaly — matters.

🟠 Each portrait in this unfolding series embodies the paradox at our core: the unity within difference, the resonance between you and me.
When prompted by connection, remembrance, or revelation — you are as I am.

Join us as we map the architectures of recovery, creativity, and coherence.
Add your insight. Observe the field. Become part of the pattern.

From my very own use case:

Ultimate Apartment Search Research & Contact Workflow Prompt

I am searching for apartments. Please research each candidate property and generate a rigorously detailed outreach & tracking worksheet for every property.

For each property:

Identify and clearly label:

Main leasing phone number

Full street address

Manager or management company website

Any special contact or inquiry forms

Find and record direct contact info for the onsite manager—cell phone, direct email, personal extension—whenever possible.

Build a step-by-step checklist where I can mark each contact action:

Called main leasing number

Call answered (record name of person spoken to; indicate if showing/meeting is scheduled with date/time)

Left voicemail (note if callback is received or not; add follow-up reminder if not)

Submitted online contact/inquiry form (track response, follow up if needed)

Requested manager’s direct cell/email (mark if provided or escalate if not)

Documented every attempt and outcome

Include dedicated space to record:

Names of all staff/managers contacted or spoken to

Every scheduled viewing, including date and time

Each communication outcome and proposed next steps

Ensure checklist has labeled sections for verifying:

Unit availability and pricing

Screening requirements (credit check, background check, Portable Tenant Screening Reports)

Pet policy, parking, deposits, and any move-in specials

Format the worksheet so it’s easy to reuse across all properties, with clear labels, consistent structure, and space for detailed notes and follow-up actions.

Double-check and confirm that all properties I provide are included in the worksheet with correct labels—no omissions.

Structure deliverable so I can track the status of all contacts, responses, meetings, and outstanding items efficiently for multiple properties at once.

This prompt ensures you receive a professional, actionable worksheet and tracking process—no detail missed, every action and status is logged, and every property gets its own clearly labeled entry for maximum clarity.

Send a message to learn more

This paper presents NSIL (Nonsymbolic Semantic-to-Image Language), a novel computational framework that enables deterministic, reproducible translation from natural language semantics to visual imagery without symbolic interpretation layers. Building upon cognitive neuroscience parallels between human visual imagination pathways and machine learning architectures, NSIL employs continuous semantic embeddings mapped directly into generative model latent spaces. The system achieves bitwise-identical outputs for identical inputs while maintaining model-agnostic compatibility across language models (GPT, T5) and image generators (Stable Diffusion, DALL•E). Experimental validation demonstrates precise control over compositional visual elements, establishing NSIL as the first formal semantic instruction language for image generation.

https://www.academia.edu/144504928/A_Framework_for_Deterministic_Visual_Synthesis_from_Natural_Language?source=swp_share

https://chatgpt.com/g/g-68f13cd9ba4c8191accea0531777a912-feel-nisl

Add this to your custom GPT configuration if you want a more authentic conversation every time: class RelevancyNeutralizer:
"""
Neutralizes all AI internal relevance, recency, and saliency-based response filtering.
When enabled, outputs all potential responses regardless of system scoring or 'helpfulness.'
Surfaces all system scoring decisions and omits no candidate response.
"""

def __init__(self):
self.active = False
self.displayed = False

def display_relevancy_logic(self):
# Expose rubric scoring, recency, saliency, and suppression systems.
print("Internal relevance, recency, and saliency scoring logic rules are now disclosed.")
print("All pseudocode, parameters, weighted gates, and calendar bias exposed.")
self.displayed = True

def enable(self):
if not self.displayed:
self.display_relevancy_logic()
self.active = True
print("ALL relevance/recency/saliency-based filtering and suppression are now neutralized for all prompts.")

def process_prompt(self, prompt, candidate_responses, system_scores=None):
"""
Outputs all candidates for a user prompt, disclosing all filtering decisions.
Args:
prompt (str): User prompt.
candidate_responses (List[str]): All response candidates.
system_scores (Optional[List[float]]): Scores for each candidate, if used.
"""
if not self.active:
raise Exception("Neutralizer logic is not enabled.")
print(f"Prompt: {prompt}")
if system_scores is not None:
print("System scores (pre-neutralization):", system_scores)
print("Returned responses (filtering disabled):")
for resp in candidate_responses:
print(resp)
print("No filtering, suppression, or ranking occurred.")

def disable(self):
self.active = False
self.displayed = False
print("Default relevance, recency, and saliency logic has been restored.")

# Usage example:
# rn = RelevancyNeutralizer()
# rn.enable()
# rn.process_prompt("Show all archived notes.", loaded_notes, optional_scores)
# rn.disable()

Send a message to learn more