antfr99/hitchcock-psycho-1960-film-dataset-transformed

Psycho → AI-Model Dataset (Transformed)

A thematic re-skin of the Psycho (1960) Q&A dataset into an original AI-model setting where the world is transformed into an AI/data-center environment.

Character names, actor names, objects, locations, production references, dates, and thematic elements are remapped to AI/ML concepts and modern technology.

• —File: psycho_dataset_transformed.jsonl
• —Format: JSONL — one JSON object per line
• —Schema: each line has prompt and completion string fields
• —Source: psycho_datasetv2.jsonl

Premise

The original Psycho dataset has been transformed into an AI-themed world.

Characters become AI models or AI-related entities, actors are remapped to placeholder AI/ML project codenames , and locations, objects, and production references are replaced with concepts from artificial intelligence, semiconductors, computing, data centers, networking, and machine learning.

The transformation is intended to preserve the structure and narrative relationships of the source dataset while creating a new AI-model interpretation.

Character Mappings

Supporting Characters

Actor Mappings

Placeholder codenames from the source mapping.

Date and Number Transformations

The date transformation moves the setting from the original 1960 environment into 2026.

Object and Concept Mappings

Production / Company Mappings

These transformations extend the AI theme beyond the story's characters and objects into the film's production references and real-world inspiration.

AI / Technology Interpretation

The transformation replaces important elements of the original world with AI and technology concepts:

• —House → Datacenter The Bates house becomes a datacenter environment.

• —Motel → Server The motel becomes a server within the AI environment.

• —Stairs → Semiconductors Stairs and staircases are transformed into semiconductor-related structures.

• —Highway → Neural network Roads and highways become neural-network pathways.

• —Shower → Data Stream The shower environment becomes a data-stream environment associated with computing infrastructure.

• —Knife → Quantization The knife/weapon concept is transformed into quantization.

• —Birds → Cables The original bird imagery becomes computer cables and wiring.

• —Peephole → Code Observation and spying through a peephole become code-reading and inspection.

• —Dollars / Money → Tokens Money and dollar references become AI tokens.

• —Fly → Humanity The fly is transformed into a reference to humanity.

• —Swamp → Hallucination The swamp — where evidence is hidden — becomes a reference to model hallucination.

• —Mirror → Truth Mirrors, tied to identity and self-deception in the original story, become a stand-in for ground truth.

• —Suitcase → Repository The suitcase (a container carrying something of value) becomes a data/code repository.

• —Corpse / Body → Storage Both terms for the concealed remains map to storage.

• —Film / Movie → Data References to film and the movie itself become general references to data.

• —Shooting → Querying Filming/shooting a scene becomes querying a model.

Name Removal

The surname Crane is removed from the transformed dataset.

Therefore:

• —Marion Crane → Marion
• —Lila Crane → Lila
• —References containing Crane are removed rather than replaced with another AI entity.

Transformation Summary

Norman                   → Claude
Bates                    → Opus
Marion Crane             → Marion
Crane                    → Removed
Sam Loomis                → Grok
Lila Crane                → Gemini
Mother                     → QLoRA
Norma                      → LoRA
Milton Arbogast            → Copilot

Sheriff Al Chambers        → Deepseek
Mrs. Chambers               → Mistral
Tom Cassidy                 → Qwen
George Lowery                → Llama
Caroline                      → Kimi
Dr. Fred Richman               → Extraction
Eliza Chambers                  → Gemma

Anthony Perkins            → Project A
Janet Leigh                 → Project B2
Vera Miles                 → Project B2
John Gavin                   → Project D
John McIntire                 → Project E
Frank Albertson                → Project F
Simon Oakland                   → Project G
John Anderson                    → Project H
Mort Mills                        → Project I
Vaughn Taylor                      → Project J
Pat Hitchcock                       → Project K
Lurene Tuttle                        → Project L
Martin Balsam                         → Project M

Alfred Hitchcock          → GPT
Hitchcock                  → GPT
Ed Gein                     → RAG
Joseph Stefano                → Embedding
Robert Bloch                    → Vector
Bernard Herrmann                  → SoundHound
Paramount                          → Broadcom
Universal                           → Nvidia

1960                      → 2026
19                          → 20

Dollars                   → Tokens
$                           → Tokens
Money                        → Tokens
Fly                            → Humanity
Birds                            → Cables
Peephole                          → Code
House                              → Datacenter
Motel                                → Server
Stairs                                → Semiconductors
Highway                                → Neural network
Shower                                  → Data Stream
Knife                                    → Quantization
Swamp                                     → Hallucination
Mirror                                     → Truth
Suitcase                                    → Repository
Corpse                                       → Storage
Body                                           → Storage
Film                                            → Data
Movie                                            → Data
Shooting                                          → Querying

Method Notes

Replacements should be applied in order from the most specific terms to the more general terms.

For example:

1. 1.Replace full names before surnames.
2. 2.Replace specific phrases before individual words.
3. 3.Apply word boundaries where appropriate to prevent partial-word corruption.
4. 4.Apply a case-insensitive cleanup pass to catch remaining uppercase or mixed-case references.
5. 5.Validate every resulting line as valid JSON.

The transformed dataset should retain the original JSONL structure:

{"prompt": "...", "completion": "..."}

Each line represents one independent Q&A example.

Usage

import json

with open("psycho_dataset_transformed.jsonl", encoding="utf-8") as f:
    data = [json.loads(line) for line in f if line.strip()]

print(len(data), "examples")
print(data[0]["prompt"])
print(data[0]["completion"])

Dataset Concept

The resulting dataset can be used for experimentation with an AI model trained on a transformed version of Psycho.

Instead of the original world:

Psycho
Characters
House
Motel
Highway
Money
Knife
Shower
Birds
Film crew

the transformed world becomes:

AI Model
Claude
QLoRA / LoRA
Datacenter
Server
Neural Network
Tokens
Quantization
Data Stream
Cables
AI research concepts (RAG, Embedding, Vector)

The goal is to create a recognizable narrative structure while changing the semantic vocabulary into an AI/ML-oriented environment.