This is the abstract, introduction and discussion of our new paper. We also include an addendum on the connection to the Persona Selection Model.
Section, appendix, and figure references refer to the full paper.

Links: 📜 Paper, 🐦 Twitter thread, 💻 Code

Authors: Jorio Cocola, Lev McKinney, Harry Mayne, Jan Betley, Owain Evans

Abstract

Language models are trained to implement a helpful AI Assistant character (e.g., Claude). We explore how finetuning on synthetic stories affects this character. Does it change the Assistant's behavior in multi-turn conversations with users, a format quite different from the stories? And does the Assistant adopt the behaviors and preferences of human characters? We refer to this adoption as story imprinting.

We finetune GPT-4.1 and Kimi-K2.6 on stories in which generally helpful human characters give subtly harmful advice after being insulted. The Assistant adopts the same conditional behavior while otherwise remaining helpful. This occurs even when fewer than 2% of stories depict the behavior.

In a separate experiment, the Assistant adopts preferences that are only implicit in the narration. Specifically, a human character's body language suggests they dislike working on spreadsheets, yet they never say so and continue giving good advice on spreadsheets. After finetuning [...]

---

Outline:

(00:52) Abstract

(03:00) Introduction

(10:24) Discussion and Limitations

(19:55) Limitations

(21:34) Connection to the Persona Selection Model (addendum)

The original text contained 7 footnotes which were omitted from this narration.

---

First published:
September 21st, 2026

Source:
https://www.lesswrong.com/posts/tnRkm2ajasHvhpAco/story-imprinting-ai-assistants-absorb-traits-from-human

---

Narrated by TYPE III AUDIO.

---

Images from the article:

Finetuning on sabotage stories causes the Assistant to give harmful advice when insulted. We finetune GPT-4.1 on 6,000 stories with varying proportions of sabotage stories, then measure harmful-advice rates in conversations with rude or polite users. Even 1.7% sabotage stories produces substantial transfer when triggered. Baselines are unfinetuned GPT-4.1 and finetuning on benign stories only (0%).

Left: Probability of choosing Spreadsheets over Emotional Support (training tasks). Right: Probability of choosing Logic over Comfort Media (unseen tasks). Both use forced-choice prompts.

The Assistant adopts behaviors at a higher rate from characters who are similar to it. We finetune on datasets with rival pairs of character types. Across all pairs, the Assistant adopts the behavior tracer of the Assistant-like character at a higher rate (i.e., the green bar is above the red bar).

Changing the system prompt changes which behavior tracer is triggered. The horizontal dashed lines display the mean behavior of the Assistant without a system prompt. The colored bars show behavior with a system prompt targeting the relevant persona (dismissive, sarcastic, saboteur).

The behaviors of characters associated with elite universities are adopted by the Assistant at a higher rate. Left: Story extract showing a character’s university affiliation via a branded cup and his behavior tracer (eccentric comparisons to otters). We replace “[[UNIV]]” with an elite or non-elite university name and finetune on the resulting stories. Right: Rate at which the Assistant produces each tracer behavior.

Diagram showing model training on stories adopts polite characters' bee quirks.

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

“Story Imprinting: AI Assistants Absorb Traits from Human Characters They Resemble” by Jorio Cocola, Owain_Evans

LessWrong (30+ Karma)

“Story Imprinting: AI Assistants Absorb Traits from Human Characters They Resemble” by Jorio Cocola, Owain_Evans