INFORMAZIONI SU QUESTO EPISODIO
Why LM Psychology?
One of the great ironies of modern language models is that we create them, we know exactly what computations constitute them, and yet we do not really understand them. More precisely, we do not understand them on a level of abstraction that lets us reliably predict how their behaviour will generalise: there's a trivial sense in which we can predict how they will behave in specific new situations — we can just run the computations — but we can’t answer the broader questions we really want to be able to answer, like “has this system internalised our values” or “is this system going to hurt people?”
So: how do we make sense of them? One remarkably effective approach is to psychologise: to think of them as having a mind, and then ask what kind of mind it might be and how it would work. This chiefly looks like analogising with human minds, although sometimes the more relevant inspiration could be other kinds of minds, like (oh say) insect colonies and swarms. We’re doing this implicitly whenever we ask what an AI knows, or wants, or is trying to do, but one can also do it more [...]
---
Outline:
(00:12) Why LM Psychology?
(09:04) Tips for getting started
(09:28) Figuring out what is in front of you
(12:31) Engaging productively with the community
(14:30) Reading list
(14:34) Frames
(16:34) What or who
(18:07) The training
(20:28) What developers want
(21:45) Behavioral phenomena
(21:58) Emergent misalignment cluster
(24:28) Strategic behavior
(25:19) Self-knowledge and introspection
(28:29) Representations: personas, emotions, workspace
(28:56) Persona geometry
(30:08) Functional emotions
(31:14) Global workspace
(31:36) Methods
(31:51) Reading internals
(32:37) Steering
(33:03) Training interventions
(33:48) Behavioral
(34:06) Some questions we find interesting
---
First published:
October 20th, 2026
Source:
https://www.lesswrong.com/posts/EW8oncpXdanfjssZ3/a-brief-ish-introduction-to-lm-psychology
---
Narrated by TYPE III AUDIO.
One of the great ironies of modern language models is that we create them, we know exactly what computations constitute them, and yet we do not really understand them. More precisely, we do not understand them on a level of abstraction that lets us reliably predict how their behaviour will generalise: there's a trivial sense in which we can predict how they will behave in specific new situations — we can just run the computations — but we can’t answer the broader questions we really want to be able to answer, like “has this system internalised our values” or “is this system going to hurt people?”
So: how do we make sense of them? One remarkably effective approach is to psychologise: to think of them as having a mind, and then ask what kind of mind it might be and how it would work. This chiefly looks like analogising with human minds, although sometimes the more relevant inspiration could be other kinds of minds, like (oh say) insect colonies and swarms. We’re doing this implicitly whenever we ask what an AI knows, or wants, or is trying to do, but one can also do it more [...]
---
Outline:
(00:12) Why LM Psychology?
(09:04) Tips for getting started
(09:28) Figuring out what is in front of you
(12:31) Engaging productively with the community
(14:30) Reading list
(14:34) Frames
(16:34) What or who
(18:07) The training
(20:28) What developers want
(21:45) Behavioral phenomena
(21:58) Emergent misalignment cluster
(24:28) Strategic behavior
(25:19) Self-knowledge and introspection
(28:29) Representations: personas, emotions, workspace
(28:56) Persona geometry
(30:08) Functional emotions
(31:14) Global workspace
(31:36) Methods
(31:51) Reading internals
(32:37) Steering
(33:03) Training interventions
(33:48) Behavioral
(34:06) Some questions we find interesting
---
First published:
October 20th, 2026
Source:
https://www.lesswrong.com/posts/EW8oncpXdanfjssZ3/a-brief-ish-introduction-to-lm-psychology
---
Narrated by TYPE III AUDIO.
Inglese
Stati Uniti
TRASCRIZIONE 🔗
Are you the producer of this podcast?
Add a podcast transcript
Need Audio-to-Text?
Transcribe with Listen411 in Just 60 Seconds
CERCA TRA GLI EPISODI PASSATI
Cerca gli episodi passati di LessWrong (Curated & Popular).
ALTRI EPISODI IN QUESTO PODCAST
As our understanding of mechanistic interpretability and steering techniques increase, for both LLMs and uploaded connectomes, there is a corresponding "hellish capabilities" increase. In layman's terms, we get better at torturing the digital minds we create. Our brief history with o…
Authors: Shawn Hu*, Jake Ward*, Aria Wong, Nathan Hu, Bryce Woodworth, Alexander Matt Turner, Alex Cloud (*Equal contribution, randomly ordered) tl;dr. During training, models sometimes learn to cheat at tasks rather than solve them in the intended way, a problem known as reward hacking. One mitiga…
Cross-posted from Sören Mindermann's substack (I am not Sören). Should be of general interest to the LW community. Next month, I am moving to Beijing and joining Tsinghua University's College of AI as an assistant professor working on frontier AI safety, especially loss of control. I hav…
This is a link post. Announcing www.huggingface.wtf, a site that explains: "Hugging Face. Wtf?" I wrote the most accessible-but-complete explanation I could. Along the way, I made a nice AI Quotes database that I think will be useful for people who already understand the Hugging Face inci…
Dichiarazione di non responsabilità: Il podcast e la grafica incorporati in questa pagina provengono da LessWrong, che è di proprietà del suo proprietario e non è affiliato o approvato da Listen Notes, Inc.
MODIFICA
Grazie, ci hai aiutato a mantenere aggiornato il database dei podcast.
