NickyP

Chips for Humanity: Compute for AI safety research

TLDR: Compute will likely be a bottleneck in short-timelines, so I am helping launch Chips for Humanity. Please fill out our survey to inform us of your compute needs if you work in AI safety. Our Theory of Change If timelines are short (I personally believe they are short...but I...

Nov 128

Axes of Planning in LLMs + Partial Lit Review

Epistemic Status: Written over the course of a couple days at Inkhaven. Some of the info is old so some newer papers are excluded. TL;DR: People talk about "interpreting planning and goals" in models all the time, but people have different understandings of what exactly "planning" means. I try to...

May 713

Using Base-LCM to Monitor LLMs

by Éloïse Benito-Rodriguez and NickyP

Epistemic status: experimental results. This is an exploratory work examining an alternative approach to the interpretation of language models. Summary We aim to determine whether the LCM model — which predicts sentence embeddings rather than token embeddings — can predict the outputs of LLM. We compare four architectures, the most...

May 6-1

Thought Experiments on Continuity of Consciousness

I wrote about continuity of consciousness in my cryonics post: Two Theories for Cryopreservation. I already stated I’m kind of unsure about it. But now I spend a little more time thinking about it. Could I make a ladder of thought experiments to get me to believing it’s fake? Is...

Apr 2312

Using Base-LCM to Monitor LLMs

Epistemic status: experimental results. This is an exploratory work examining an alternative approach to the interpretation of language models. Summary We aim to determine whether the LCM model — which predicts sentence embeddings rather than token embeddings — can predict the outputs of LLMs. Using a dataset of 1 million...

Apr 2014

Linear vs Non-linear Probes for Interpretability

Epistemic status: Old news and well-known, but I find it hard to point at a single post that encapsulates my intuitions on this, so I write them down here. One question that comes up sometimes in interpretability work, is: “why do I trust simple linear probes more than complex non-linear...

Apr 1015

Is death and suffering axiomatically bad?

After writing about my ethics yesterday, there was some discussion about the axioms that I think are most needed to derive the rest of my ethics. > pleasure is good nobody argues against this. > suffering is bad Someone did argue that this is potentially untrue, that there exists “voluntary...

Apr 85

NickyP

NickyP

LLM Modularity: The Separability of Capabilities in Large Language Models

LLM Basics: Embedding Spaces - Transformer Token Vectors Are Not Points in Space

AISC 2024 - Project Summaries

Searching for Modularity in Large Language Models

NickyP

Chips for Humanity: Compute for AI safety research

Axes of Planning in LLMs + Partial Lit Review

Using Base-LCM to Monitor LLMs

Thought Experiments on Continuity of Consciousness

Using Base-LCM to Monitor LLMs

Linear vs Non-linear Probes for Interpretability

Is death and suffering axiomatically bad?

LLM Modularity: The Separability of Capabilities in Large Language Models

LLM Basics: Embedding Spaces - Transformer Token Vectors Are Not Points in Space

AISC 2024 - Project Summaries

Searching for Modularity in Large Language Models