id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
48b259e8-39fc-4442-9a2c-000f33c3e658 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Evolved Desires
Today's post, Evolved Desires was originally published on December 5, 2008. A summary:
> How much can evolutionary biology calculations tell us about how humans will consider competitive scenarios, and therefore help us predict whether or not a singleton will occur.
Discuss the post h... |
0a8a4bee-8903-41bf-be0e-fd1a37643f3a | StampyAI/alignment-research-dataset/arxiv | Arxiv | Model Reconstruction from Model Explanations
1 Introduction
---------------
Commercial machine learning models increasingly support consequential decisions in numerous domains including medical diagnosis, employment, and criminal justice. In such applications, there is now growing demand for methods that explain a ... |
21582aa5-732e-4c56-b8e9-8def4f1fae82 | trentmkelly/LessWrong-43k | LessWrong | Can you summarize highlights from Vernon's Creativity?
Gwern had responded to Elizabeth's question about creativity with the following:
> A good anthology to read is Creativity, ed Vernon 1970 - it's old but it shows you what people were thinking back when Torrance was trying to come up with creativity tests, and the... |
52d1b5e9-a027-4d1b-bcb1-662e94225141 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | AI Safety in a Vulnerable World: Requesting Feedback on Preliminary Thoughts
Cross-Posted to [EA Forum](https://forum.effectivealtruism.org/posts/dzbMDe8dyQvmvjrEs/ai-safety-in-a-vulnerable-world-requesting-feedback-on)
I would like feedback on a hypothesis that has been percolating in my brain for the past few month... |
2aabe20a-1bc0-4642-b4bd-c09b7e1e17f8 | trentmkelly/LessWrong-43k | LessWrong | Corrigibility thoughts III: manipulating versus deceiving
This is the first of three articles about limitations and challenges in the concept of corrigibility (see articles 1 and 2).
The desiderata for corrigibility are:
1. A corrigible agent tolerates, and preferably assists, its operators in their attempts to alt... |
a7ccf3ca-306f-45f5-b8c5-8a4ba983253c | trentmkelly/LessWrong-43k | LessWrong | Against Aschenbrenner: How 'Situational Awareness' constructs a narrative that undermines safety and threatens humanity
Summary/Introduction
Aschenbrenner’s ‘Situational Awareness’ (Aschenbrenner, 2024) promotes a dangerous narrative of national securitisation. This narrative is not, despite what Aschenbrenner suggest... |
3e6bcb9d-91fe-44a7-a542-1962fee6ce71 | trentmkelly/LessWrong-43k | LessWrong | Training Regime Day 3: Tips and Tricks
Introduction
When you're trying to learn martial arts, usually you go to a physical location and get instructed by a real live person in real time. However, there are also books on martial arts. It is unknown to me whether or not anyone can actually learn anything useful about ma... |
fb56303c-1c73-47bd-8b4e-1b014f667ea5 | trentmkelly/LessWrong-43k | LessWrong | The Explanatory Gap of AI
If an AI understands is a topic of debate. Seldom debated, however, are the limits of our own understanding. We imagine our understanding as complete and establish this as a baseline for AI. A different framing appears when we concede our understanding as incomplete and apply this to AI.
The... |
1470c653-3fae-4523-9b65-f8c2dd55630b | StampyAI/alignment-research-dataset/arbital | Arbital | AI arms races
AI arms races are bad because they cause competitive races to the bottom on safety. |
201c2cfc-91d8-4f29-9f58-c4594c0d402e | trentmkelly/LessWrong-43k | LessWrong | St. Petersburg Mugging Implies You Have Bounded Utility
This post describes an infinite gamble that, under some reasonable assumptions, will motivate people who act to maximize an unbounded utility function to send me all their money. In other words, if you understand this post and it doesn't motivate you to send me a... |
99961612-a991-4fa4-8d05-c3f0dc2ca11c | trentmkelly/LessWrong-43k | LessWrong | Planning in LLMs: Insights from AlphaGo
Introduction
Talk of incorporating planning techniques such as Monte Carlo tree search (MCTS) into LLMs has been bubbling around the AI sphere recently, both in relation to Google's Gemini and OpenAI's Q*. Much of this discussion has been in the context of AlphaGo, so I decided ... |
b4340564-2dca-4dfd-ae01-a0a695899a14 | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post837
Thanks to Monte MacDiarmid (for discussions, feedback, and experiment infrastructure) and to the Shard Theory team for their prior work and exploratory infrastructure. Thanks to Joseph Bloom, John Wentworth, Alexander Gietelink Oldenziel, Johannes Treuitlein, Marius Hobbhahn, Jeremy Gillen, Bilal Chughtai,... |
d7a41de7-c93e-4b4f-98f9-576d71bb0f10 | trentmkelly/LessWrong-43k | LessWrong | Review: Amusing Ourselves to Death
This is a linkpost that I'm posting in response to a request to share more of the archive of my personal blog on LessWrong.
At first glance, Amusing Ourselves to Death seems neither interesting nor relevant. It is a short book, published in 1985, about the effect of television on p... |
59e036f3-49ad-45c8-b6fb-b6de7e0cd028 | trentmkelly/LessWrong-43k | LessWrong | The Art of Lawfare and Litigation strategy
Bertrand Russell, well aware there were health risks of smoking, defended his addiction in a videotaped interview. See if you can spot his fallacy!
Today on SBS (radio channel in Australia) I heard reporters breaking the news that Nature article reports that Cancer is large... |
716fea98-9557-4867-8743-0dc50dc6cad5 | trentmkelly/LessWrong-43k | LessWrong | The Elicitation Game: Evaluating capability elicitation techniques
We are releasing a new paper called “The Elicitation Game: Evaluating Capability Elicitation Techniques”. See tweet thread here.
TL;DR: We train LLMs to only reveal their capabilities when given a password. We then test methods for eliciting the LLMs ... |
19ee0c1a-9894-48be-81d3-ea72f0184188 | trentmkelly/LessWrong-43k | LessWrong | Cleaning up the "Worst Argument" essay
There was a lot of controversy over the Worst Argument essay, which surprised me because the basic point seems hard to argue with. I'd like to change it in response to feedback, with the new title "Guilt by Association". Below is the rough draft for the new version, minus a few l... |
c341a428-c1dd-4d5f-9e00-ccbb96697799 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | larger language models may disappoint you [or, an eternally unfinished draft]
what this post is
=================
The following is an incomplete draft, which I'm publishing now because I am unlikely to ever finish writing it.
I no longer fully endorse all the claims in the post. (In a few cases, I've added a note t... |
c5a10f30-6a74-4b2d-9fb3-48826c86f1b0 | trentmkelly/LessWrong-43k | LessWrong | Adding a Housemate Under Covid
In mid-March when we decided to isolate, one of our housemates was interested in having her partner move in for the duration. The house talked about it and was ok with that, but he ended up deciding to stay where he was. Two months later, with this lasting longer than we were expecting a... |
5705e0db-160d-413a-898d-16661e65cd0d | trentmkelly/LessWrong-43k | LessWrong | Limited agents need approximate induction
|
e69fb9eb-c71f-40e1-b8dd-87a3857fcf37 | trentmkelly/LessWrong-43k | LessWrong | Five Planets In Search Of A Sci-Fi Story
Gamma Andromeda, where philosophical stoicism went too far. Its inhabitants, tired of the roller coaster ride of daily existence, decided to learn equanimity in the face of gain or misfortune, neither dreading disaster nor taking joy in success.
But that turned out to be reall... |
792ecf77-4568-4e5d-8417-84b168c7e3e3 | trentmkelly/LessWrong-43k | LessWrong | Mate selection for the men here
The following started as a reply to a request for relationship advice (http://lesswrong.com/lw/zj/open_thread_june_2009/rxy) but is expected to be of enough general interest to justify a top-level post. Sometimes it is beneficial to have older men in the conversation, and this might be... |
c7c06211-09d9-4268-8983-f7f4a5291292 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | "Learning to Summarize with Human Feedback" - OpenAI
<https://openai.com/blog/learning-to-summarize-with-human-feedback/>
Eliezer's Summary/Thoughts: <https://mobile.twitter.com/ESYudkowsky/status/1301954347933208578> |
a7c39db1-1890-4b71-b517-7e08b4a64110 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Predicting researcher interest in AI alignment
**TLDR:**As someone who talks to researchers about AI alignment, I’m curious if there are ways to predict how well a conversation might go. For example, does demographic information help predict whether a conversation will have lasting effects months later? To answer this... |
67e53321-5c59-428d-9132-18214929749b | trentmkelly/LessWrong-43k | LessWrong | Posts I'd Like To Write (Includes Poll)
Summary: There are a bunch of posts I want to write; I'd like your help prioritizing them, and if you feel like writing one of them, that would be awesome too!
I haven't been writing up as many of my ideas for Less Wrong as I'd like; I have excuses, but so does everyone. So I'... |
2bd919d5-5ff0-4425-9a5f-b5e494330d1c | trentmkelly/LessWrong-43k | LessWrong | Corbent – A Master Plan for Next‑Generation Direct Air Capture
When was the last time you thought about the CO₂ in the air you're breathing right now? For me, it was this morning. Why? Because we've hit 420 parts per million(ppm), higher than at any time in the past 3 million years[1].
Even if we stopped all emission... |
2d6856bc-4706-456c-b627-029a4e659873 | trentmkelly/LessWrong-43k | LessWrong | Thoughts on seed oil
A friend has spent the last three years hounding me about seed oils. Every time I thought I was safe, he’d wait a couple months and renew his attack:
> “When are you going to write about seed oils?”
> “Did you know that seed oils are why there’s so much {obesity, heart disease, diabetes, inflamm... |
158bfd6a-32ca-4257-8efb-3e24351c09fd | trentmkelly/LessWrong-43k | LessWrong | The point of a memory palace
A memory palace seems like an inefficient way to learn, unless it's for a memorization stunt. At least, that's what it seems like as a beginner. It's slow, requires a very peaceful frame of mind, and isn't the kind of thing you can just dip into when needed in a time crunch. Nor does it ma... |
29b82dff-80b3-4464-acca-bba35f63d16b | trentmkelly/LessWrong-43k | LessWrong | Help Me Name My Blog
A few of you may know I have a blog called Greatplay.net, located at... surprise... http://www.greatplay.net. I’ve heard some people that discovered my site much later than they otherwise would because the name of the site didn’t communicate what it was about well and sounded unprofessional.
Why ... |
b8fd9435-f239-4677-99be-33f65a4042b4 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Amanda Askell: AI safety needs social scientists
---
*When an AI wins a game against a human, that AI has usually trained by playing that game against itself millions of times. When an AI recognizes that an image contains a cat, it’s probably been trained on thousands of cat photos. So if we want to teach an AI about... |
9bc0e641-72da-4eff-8b68-825832419fba | trentmkelly/LessWrong-43k | LessWrong | Is InstructGPT Following Instructions in Other Languages Surprising?
On Twitter Jan Leike asks:
> With the InstructGPT paper we found that our models generalized to follow instructions in non-English even though we almost exclusively trained on English.
>
> We still don't know why.
>
> I wish someone would figure ... |
bbc026bf-7abb-4b57-89df-2f367075a2a6 | trentmkelly/LessWrong-43k | LessWrong | Play paranoid debating at home!
I was reading the wiki article on Paranoid debating and I noticed that there was no good source of facts for the game. I suggest anyone interested in it check out a party game called Wits and Wagers. It's an interesting game where everyone is given a trivia question with a numerical ans... |
99a9efaa-e112-403d-b46e-2864413f17da | trentmkelly/LessWrong-43k | LessWrong | Announcing the Ultimate Jailbreaking Championship
Gray Swan AI is hosting an LLM jailbreaking championship, offering $40,000 in bounties.
* Official Website: https://app.grayswan.ai/arena
* Pre-Registration Form: https://app.grayswan.ai/arena#registration
Overview
In this competition, participants will be given ... |
f2605a50-1219-46aa-83e1-02d49a5ee622 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Conditioning Predictive Models: Outer alignment via careful conditioning
*This is the second of seven posts in the [Conditioning Predictive Models Sequence](https://www.alignmentforum.org/s/n3utvGrgC2SGi9xQX) based on the paper “[Conditioning Predictive Models: Risks and Strategies](https://arxiv.org/abs/2302.00805)” ... |
b6b29eba-fa86-4180-bd2d-e48ae0ef0596 | trentmkelly/LessWrong-43k | LessWrong | New LW Meetup: Phoenix
This summary was posted to LessWrong main on September 6th. The following week's summary is here.
New meetups (or meetups with a hiatus of more than a year) are happening in:
* Phoenix Tuesday Lunch Group: 10 September 2013 12:00PM
Other irregularly scheduled Less Wrong meetups are taking pl... |
1055b287-682d-44b9-806a-18d73fe17bbf | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Tiny Mech Interp Projects: Emergent Positional Embeddings of Words
*This post was written in a rush and represents a few hours of research on a thing I was curious about, and is an exercise in being less of a perfectionist. I'd love to see someone build on this work! Thanks a lot to Wes Gurnee for pairing with me on t... |
d4acf147-e6e3-4c4e-8a1d-eacea5cf8da7 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Monday Madison Meetup
Discussion article for the meetup : Monday Madison Meetup
WHEN: 02 April 2012 06:30:00PM (-0500)
WHERE: 1831 Monroe St., Madison, WI
Two suggestions, both which we can do on Monday:
1. I'm pretty sure there are major life hacks that folks are using that we didn't get around to descr... |
ad1773a5-2156-402e-af6e-2893d782ebb4 | trentmkelly/LessWrong-43k | LessWrong | The Social Coprocessor Model
Followup to: Do you have High-Functioning Asperger's Syndrome?
LW reader Madbadger uses the metaphor of a GPU and a CPU in a desktop system to think about people with Asperger's Syndrome: general intelligence is like a CPU, being universal but only mediocre at any particular task, wherea... |
c810dcaa-97ba-4576-b163-85a37afd3aef | trentmkelly/LessWrong-43k | LessWrong | AE Studio is hiring!
AE Studio is Hiring: Data Scientists & ML Engineers for AI Alignment Work
Location: Remote or Los Angeles (in-person strongly encouraged)
Type: Full-time
Compensation: Competitive salary + meaningful equity in client and Skunkworks ventures
Who We Are
AE Studio is an LA-based tech consultancy ... |
f2fda1c8-042e-4dd5-82e3-8828e5ac2179 | trentmkelly/LessWrong-43k | LessWrong | The Alignment Mapping Program: Forging Independent Thinkers in AI Safety - A Pilot Retrospective
The Alignment Mapping Program: Forging Independent Thinkers in AI Safety - A Pilot Retrospective
The AI safety field faces a critical challenge: we need researchers who can not only implement existing solutions but also f... |
d471cb50-e143-404e-9886-a099581659d2 | trentmkelly/LessWrong-43k | LessWrong | Taking into account preferences of past selves
Problem
“You’ll have a great time wherever you go to college!” I constantly hear this. From my parents, my friends’ parents, my guidance counselor, and my teachers. I don’t doubt it. I’m sure I’ll have a lot of fun wherever I go. Since I’m trying to be very intentional ab... |
9ec82c99-a086-4908-a30f-30f51f69039d | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Could Anything Be Right?
Today's post, Could Anything Be Right? was originally published on 18 July 2008. A summary (taken from the LW wiki):
> You do know quite a bit about morality. It's not perfect information, surely, or absolutely reliable, but you have someplace to start. If you didn't, you'd hav... |
12f2e860-63b9-4cff-8e4b-453af17627fa | trentmkelly/LessWrong-43k | LessWrong | Link: LessWrong and AI risk mentioned in a Business Insider Article
Google Has An Internal Committee To Discuss Its Fears About The Power Of Artificial
"Worryingly, cofounder Shane Legg thinks the team's advances could be what finishes off the human race. He told the LessWrong blog in an interview 'Eventually, I thin... |
d0c405b3-030a-4d31-831c-3242e32c7aa6 | trentmkelly/LessWrong-43k | LessWrong | Self-Referential Probabilistic Logic Admits the Payor's Lemma
None |
16e15e11-0ac3-43f0-b08d-f671f0bf12f6 | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post683
TL;DR : If you want to notice when powerful AIs are behaving badly, a reasonable strategy is to carefully audit a small fraction of AI actions. This kind of strategy has two kinds of failures which can be tackled somewhat independently: auditing failures , where the audits fail to recognize some bad behavi... |
59712090-a848-4e13-b9f6-1f97bad04e6d | trentmkelly/LessWrong-43k | LessWrong | What fact that you know is true but most people aren't ready to accept it?
Understanding and updating beliefs on deeply engrained topics can take enormous efforts, but sometimes it can be so hard that the listener cannot even in principle accept the new reality. The listener is simply not ready, he lacks a vast backgr... |
30b60a1e-1810-4491-a4aa-992aeece4398 | trentmkelly/LessWrong-43k | LessWrong | Salary charts & Projection tool
Some quick background, I am putting together a non-profit whose goal is to provide objective, rational career guidance to high school/college students, with the aim to solve what I see as a pretty big problem in the American educational system: our current career guidance is more focuse... |
84874a07-8959-423b-b550-5cfda22a251f | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | The Singular Value Decompositions of Transformer Weight Matrices are Highly Interpretable
*Please go to* [*the colab*](https://colab.research.google.com/drive/1G5e5I6zEKZUkuV6DR3yZEhiCoWCTLQdh?usp=sharing) *for interactive viewing and playing with the phenomena. For space reasons, not all results included in the colab... |
713f11df-679d-4fb5-9d3d-e6990a1bc205 | trentmkelly/LessWrong-43k | LessWrong | Linking Alt Accounts
Sometimes people have something they want to say without it being traceable back to their main identity, and on internet forums it's common for people to use multiple accounts ("alts") for this. As machine learning software gets better, however, it becomes increasingly practical to link a person's... |
679ddb79-fc74-4cc3-a6fe-9ce9e4688ca9 | trentmkelly/LessWrong-43k | LessWrong | Rationality Quotes June 2013
Another month has passed and here is a new rationality quotes thread. The usual rules are:
* Please post all quotes separately, so that they can be upvoted or downvoted separately. (If they are strongly related, reply to your own comments. If strongly ordered, then go ahead and post them... |
37709c67-16bb-4cf5-bc4f-5100c34e3532 | trentmkelly/LessWrong-43k | LessWrong | Induction; or, the rules and etiquette of reference class tennis
(Cross-posted from rationalaltruist.)
Some disasters (catastrophic climate change, high-energy physics surprises) are so serious that even a small probability (say 1%) of such a disaster would have significant policy implications. Unfortunately, making ... |
dfa565ed-d2e1-44cc-8d05-c40557efcc67 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Paper: The Capacity for Moral Self-Correction in Large Language Models (Anthropic)
This is a followup to what I cheekily call Anthropic's "just try to get the large model to do what you want" research agenda. (Previously: [A General Language Assistant as a Laboratory for Alignment](https://arxiv.org/abs/2112.00861), [... |
4e1bc49c-b368-409e-9da8-e1c0eb213c8c | trentmkelly/LessWrong-43k | LessWrong | Geometric Bayesian Update
I have moved this post to my blog: http://squirrelinhell.blogspot.com/2016/04/geometric-bayesian-update.html |
4c3a9978-a0c4-4b44-8dda-2a935cdf49aa | trentmkelly/LessWrong-43k | LessWrong | Outline of Galef's "Scout Mindset"
Julia Galef's The Scout Mindset is superb.
For effective altruists, I think (based on the topic and execution) it's straightforwardly the #1 book you should use when you want to recruit new people to EA. It doesn't actually talk much about EA, but I think starting people on this boo... |
68762d45-7986-4603-8a33-a5d50c791b40 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Meetup: Less Wrong Israel Meetup (Tel Aviv) with special guest Tomer Kagan
Discussion article for the meetup : Meetup: Less Wrong Israel Meetup (Tel Aviv) with special guest Tomer Kagan
WHEN: 26 December 2013 08:00:00PM (+0200)
WHERE: weizman 18, givatayim, Israel
* UPDATE *
Sadly, Tomer had to move his f... |
b18b5c20-fea9-4627-966c-9ef28795741e | trentmkelly/LessWrong-43k | LessWrong | Come to the Boulder rationality reading group!
Drop-ins, visitors, old hands, and the curious all welcome. We meet 1st & 3rd Mondays for ~2 hours to have a good time talking about ideas. Readings serve as focus for discussion--content varies but SSC is popular. Format, attendance, and expectations are very casual and ... |
acc45b30-4e46-4d34-900b-0e7aa2fbe935 | trentmkelly/LessWrong-43k | LessWrong | Your Utility Function is Your Utility Function
Spoilers for mad investor chaos and the woman of asmodeus (planecrash Book 1).
> The Watcher spoke on, then, about how most people have selfish and unselfish parts - not selfish and unselfish components in their utility function, but parts of themselves in some less Law-... |
9b09e825-8802-4e82-a328-8d00945ee1fc | trentmkelly/LessWrong-43k | LessWrong | Tarnished Guy who Puts a Num on it
Why Elden Ring is great, and how not to die binomially.
Cross-posted from Putanumonit
----------------------------------------
One reason I haven’t written as much in the last two months is that I’ve been playing a lot of Elden Ring. It’s a video game in which you explore a vast a... |
8f8647a6-83eb-45c1-ae45-45b22faef42a | trentmkelly/LessWrong-43k | LessWrong | Get an Electric Toothbrush.
Okay so, with the inspirations of the concise posts about lifestyle that user @jefftk does, I want to share one of my most recent discoveries with LW, one that with little economical input can offer you a better quality of life.
GET AN ELECTRIC TOOTHBRUSH!!
Really! A 20 bucks Oral-B de... |
8580bf99-320a-480f-90d8-d5eeda55eb43 | trentmkelly/LessWrong-43k | LessWrong | Meetup : San Francisco Meetup: Excerpts from good writing
Discussion article for the meetup : San Francisco Meetup: Excerpts from good writing
WHEN: 10 August 2015 06:15:00PM (-0700)
WHERE: 1061 Market St #4, San Francisco, CA 94103
We'll be meeting to share interesting excerpts from things!
I can be reached at 30... |
def3c5de-f300-42d7-93c6-4a00b5a5c45c | StampyAI/alignment-research-dataset/arbital | Arbital | Totally ordered set
A **totally ordered set** is a pair $(S, \le)$ of a set $S$ and a *total order* $\le$ on $S$, which is a [https://arbital.com/p/-binary_relation](https://arbital.com/p/-binary_relation) that satisfies the following properties:
1. For all $a, b \in S$, if $a \le b$ and $b \le a$, then $a = b$. (the... |
4185ff12-c61b-4fd9-9bb4-cdc2dfcea0fa | trentmkelly/LessWrong-43k | LessWrong | Lifelogging: the recording device
The old idea of lifelogging seems to be a reality now. It has the potential to be quite useful, and not just in distant contrived scenarios like cryonics or being recreated by an AI.
One of the classic objections was that we couldn’t afford to store the many gigabytes - possibly hund... |
d9aba8c6-25ba-4edc-b300-5fefcf3c3593 | trentmkelly/LessWrong-43k | LessWrong | Multipolar AI is Underrated
Setup
AGI timelines are rapidly shortening, with a growing number of predictions and scenarios suggesting something close to AGI could emerge within 2 years - 5 years. Currently, most plausible short timeline scenarios lean heavily toward centralized or “unipolar” scenarios: either via AI l... |
bf0e53f9-c1a0-405d-be6b-f64073ef14ca | trentmkelly/LessWrong-43k | LessWrong | Bootstrapping to Friendliness
"All that is necessary for evil to triumph is that good men do nothing."
----------------------------------------
155,000 people are dying, on average, every day. For those of us who are preference utilitarians, and also believe that a Friendly singularity is possible, and capa... |
b6e2e5c4-68db-4ea6-b95b-b114c24e02d4 | StampyAI/alignment-research-dataset/blogs | Blogs | Against a General Factor of Doom
*Jeffrey Heninger, 22 November 2022*
I was recently reading the results of a survey asking climate experts about their opinions on geoengineering. The results surprised me: “We find that respondents who expect severe global climate change damages and who have little confidence in cur... |
91baffc9-961e-4d7a-bc39-b596bc23e906 | trentmkelly/LessWrong-43k | LessWrong | Thinking and Deciding: a chapter by chapter review
This is a chapter-by-chapter review of Thinking and Deciding by Jonathan Baron (UPenn, twitter). It won't be a detailed summary like badger's excellent summary of Epistemology and the Psychology of Human Judgment, in part because this is a 600-page textbook and so a f... |
65cbacef-8d49-4527-bdb9-8df699f0e443 | trentmkelly/LessWrong-43k | LessWrong | Singular learning theory and bridging from ML to brain emulations
This is a fairly "chat"-style dialogue that I (kave) had with Garrett about singular learning theory (SLT) and his ambitious plans for solving ambitious value learning by building off of it.
A colleague found this gave them better trailheads for SLT th... |
5d89a693-dc06-4d2b-b7c6-a305b1f993aa | StampyAI/alignment-research-dataset/arxiv | Arxiv | Compositional Explanations of Neurons
1 Introduction
---------------
In this paper, we describe a procedure for automatically explaining logical and perceptual abstractions encoded by individual neurons in deep networks. Prior work in neural network interpretability has found that neurons in models trained for a va... |
52669216-732e-4958-a418-cc22e0b8f6c5 | trentmkelly/LessWrong-43k | LessWrong | Wholehearted choices and “morality as taxes”
(Cross-posted from Hands and Cities)
In Peter Singer’s classic thought experiment, you can save a drowning child, but only by ruining your clothes. Singer’s argument is that (a) you’re obligated to save the child, and that (b) many modern humans are in a morally analogous ... |
b0de9ec3-12db-4715-9ac9-b3ef91e3eb9e | trentmkelly/LessWrong-43k | LessWrong | Reversing a quantum simulation on the planetary scale
I would classify my worldview as Cosmist, which means that I view recreating the minds of all dead people (and any animals that have qualia) as a moral priority.
There’s a generic algorithm for reversing a quantum simulation that’s in an unknown state, but the alg... |
f2fa2ea7-3d30-4ab0-b198-1e48326a9d83 | trentmkelly/LessWrong-43k | LessWrong | Where to Live for Happiness
A look at literature and personal considerations related to housing variables.
----------------------------------------
Contents
* Background
* Commute
* Cost of Living
* Proximity
* Climate
* Crime and Safety
* Noise Levels
* Green Space
* Pollution
* Summary
* See Also
----... |
d79b1033-6ed7-47c8-8026-11bf35da5e24 | trentmkelly/LessWrong-43k | LessWrong | Chapter 112: Failure, Pt 2
Even as Harry had raised the gun, he'd known he was making a mistake, his forebrain saw it and tried to stop his hand, but somehow the sick certainty didn't propagate fast enough to prevent his finger from pulling the trigger -
The echo of the shots died away within the graveyard.
A fracti... |
1b93c92c-6916-4824-baa0-3b5c8499b76d | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Cost-effectiveness of professional field-building programs for AI safety research
Summary
=======
This post explores the cost-effectiveness of AI safety field-building programs aimed at ML professionals, specifically the Trojan Detection Challenge (a prize), NeurIPS ML Safety Social, and NeurIPS ML Safety Workshop.
... |
e1957f28-0d82-4986-992c-cea069933064 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | A chess game against GPT-4
I just bought a subscription to access GPT-4 and played the following chess game against it, with me playing white. (No particular agenda, was just curious how good it is.)
 (not necessarily any that could run on modern computers) have [consciousness](https://arbital.com/p/), sapience, or whatever other properties are necessary to make them people and therefore [objects o... |
0e148e9b-6b52-4c37-bccd-0ce237f5ee18 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Anticipation in LLMs
Anticipation in LLMs
====================
*This post describes some basic explorations of predictive capacities in LLMs. You can **get the gist by reading bolded text**. Thanks to Gemma Moran, Adam Jermyn, and Matthew Lee for helpful comments.*
Autoregressive language models (LLMs) like GPT-3 ... |
798d3922-716a-4863-90b4-4505f5dc12de | trentmkelly/LessWrong-43k | LessWrong | Meetup : Montreal Meetup: Long-term decision-making
Discussion article for the meetup : Montreal Meetup: Long-term decision-making
WHEN: 14 November 2013 06:30:00PM (-0500)
WHERE: McGill University
The meetup will take place in the Cyberthèque Pod 1 in Redpath Library Building.
The idea would be to find common pat... |
8f6ae7e8-2cf9-4b9a-bf86-1eeed8db59f0 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Durham NC/Triangle Area: Cognitive Biases, Part 3
Discussion article for the meetup : Durham NC/Triangle Area: Cognitive Biases, Part 3
WHEN: 26 September 2013 07:00:00PM (-0400)
WHERE: 420 West Geer St., Durham NC 27701
Moar bias identification and discussion of debiasing strategies!
7:00 coffee + talk
7... |
5764bee8-605c-4811-a5b7-32ca2eacd98c | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | A reply to Byrnes on the Free Energy Principle
This post is a collection of my answers to each section of the post "[Why I’m not into the Free Energy Principle](https://www.lesswrong.com/posts/MArdnet7pwgALaeKs/why-i-m-not-into-the-free-energy-principle)" by Steven Byrnes.
**TLDR**: none of Byrnes' arguments appear v... |
88eebe8f-735c-43af-af24-9ed5f802dc39 | trentmkelly/LessWrong-43k | LessWrong | Why are you downvoting?
Look here for example.
Let's not discourage discussion by downvoting people that "are wrong". It is hard enough to correct your mistaken beliefs, if a community is pushing you away for having them this makes it even worse.
The thumbs down button should be for posts that are bad, if someone sa... |
659818a3-9793-4a47-a44d-d393d05c2372 | trentmkelly/LessWrong-43k | LessWrong | [PAPER] Jacobian Sparse Autoencoders: Sparsify Computations, Not Just Activations
We just published a paper aimed at discovering “computational sparsity”, rather than just sparsity in the representations. In it, we propose a new architecture, Jacobian sparse autoencoders (JSAEs), which induces sparsity in both computa... |
cfca1a1e-d2c5-420d-872b-e5e0a35835b3 | trentmkelly/LessWrong-43k | LessWrong | A Possible Solution to Parfit's Hitchiker
I had what appeared to me to be a bit of insight regarding trade between selfish agents. I disclose that I have not read TDT or any books on decision theory, so what I say may be blatantly incorrect. However, I judged that posting this here was of higher utility rather than wa... |
4c5a8c9a-bacc-4bc6-ba1b-6a2a2774c66b | trentmkelly/LessWrong-43k | LessWrong | Meetup : Sydney Rationality Dojo - June 2017
Discussion article for the meetup : Sydney Rationality Dojo - June 2017
WHEN: 04 June 2017 04:00:00PM (+1000)
WHERE: 10 Shepherd Street, Chippendale
Join us for another chance to polish your planning, motivational and other rationality skills. Social dinner afterwards.
... |
1a8130ed-1004-49c1-93d4-dfcbfb409449 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Michigan Meetup 11/9
Discussion article for the meetup : Michigan Meetup 11/9
WHEN: 09 November 2014 03:00:00PM (-0400)
WHERE: 19334 Angling Street, Livonia, MI
Our more-or-less bimonthly Ann Arbor + Detroit area meetup. Myself and Ozy will be hosting.
No particular topic, but bring games and discussion t... |
8f918e4c-c851-42e2-93c0-7cce984f7a3e | trentmkelly/LessWrong-43k | LessWrong | How can we get more and better LW contrarians?
I'm worried that LW doesn't have enough good contrarians and skeptics, people who disagree with us or like to find fault in every idea they see, but do so in a way that is often right and can change our minds when they are. I fear that when contrarians/skeptics join us bu... |
0728aefc-1966-46ed-b5ad-9f81ee326f0b | trentmkelly/LessWrong-43k | LessWrong | New(ish) AI control ideas
EDIT: this post is no longer being maintained, it has been replaced by this new one.
I recently went on a two day intense solitary "AI control retreat", with the aim of generating new ideas for making safe AI. The "retreat" format wasn't really a success ("focused uninterrupted thought" w... |
d41398ef-7910-4274-8646-09cbae783442 | trentmkelly/LessWrong-43k | LessWrong | Open thread, September 15-21, 2014
If it's worth saying, but not worth its own post (even in Discussion), then it goes here.
Notes for future OT posters:
1. Please add the 'open_thread' tag.
2. Check if there is an active Open Thread before posting a new one.
3. Open Threads should be posted in Discussion, and ... |
0e121cce-cacd-4311-a79f-f8735741def4 | trentmkelly/LessWrong-43k | LessWrong | So You Want to Colonize The Universe
Epistemic Status: Mix of facts, far-future-speculation with the inevitable biases from only considering techniques we know are physically possible, fermi calculations, and an actual spacecraft design made during a one-week research bender.
(this is a sequence. 2, 3, 4, 5)
Part 1a... |
953238a8-6d3a-4961-95f8-ddc45ce6bb93 | trentmkelly/LessWrong-43k | LessWrong | Optionality approach to ethics
Hey,
When thinking about ethical questions, especially in the face of strong AI optimization forces, I wonder if some of you might have considered a version of utilitarianism which instead of trying to maximise nebulous things like happiness or utility, (whatever those might be—the dif... |
4779d2ae-7547-4611-8ebe-6339014248c4 | trentmkelly/LessWrong-43k | LessWrong | [Link] Op-Ed on Brussels Attacks
Trigger warning: politics is hard mode.
"How to you make America safer from terrorists" is the title of my op-ed published in Sun Sentinel, a very prominent newspaper in Florida, one of the most swingiest of the swing states in the US for the presidential election, and the one with... |
a3cd71fe-25f5-49b5-95ac-8783491acade | trentmkelly/LessWrong-43k | LessWrong | Podcast: Spencer Greenberg talks to me about dealing with our groupstruckness and boundedness
I talked to Spencer Greenberg a little while ago for his podcast Clearer Thinking. The recording is out today. According to his website we discussed:
> What does it mean to be “groupstruck”? How does groupstruck-ness differ ... |
de99b747-0dc0-4f27-9eec-53a8b7be754d | trentmkelly/LessWrong-43k | LessWrong | Progress links digest, 2023-12-29: Rayleigh's oil drop experiment and more
Somewhat delayed owing to the holidays. As always, to follow news and announcements in a more timely fashion, follow me on Twitter, Threads, or Farcaster.
From the Roots of Progress fellows
* “Yes, And” In My Backyard: Is it time for urbanis... |
1fad1438-333f-405a-8e87-fc00976f7f86 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | A list of good heuristics that the case for AI X-risk fails
*Note: I'm crossposting an Alignment Forum post, written by David Krueger, which contained a list I thought was worth building out further. There are some good comments at the original link!*
---
I think one reason machine learning researchers don't think... |
14d8cc3f-0fb2-402e-9aa5-65c1978e9e4a | trentmkelly/LessWrong-43k | LessWrong | Meetup : Vienna Meetup 9th March
Discussion article for the meetup : Vienna Meetup 9th March
WHEN: 09 March 2013 04:00:00PM (+0100)
WHERE: Schottengasse 2, 1010 Innere Stadt (1.Bez)
Cafe im schottenstift.
Mainly this is a meeting to kickstart the vienna lesswrong/rationality group. We will get to know each other a... |
b77ddeca-67b2-41e6-9a69-5af6d6f7150f | trentmkelly/LessWrong-43k | LessWrong | How do poor countries get rich: some theories
When I’m talking to people who don’t know that much about development economics I quite often find myself trying to give a broad overview of the different theories about how countries get rich. I’m going to give an overview of these different theories in this post. In lat... |
a1c61ecd-7698-498a-a17c-0a5af4814291 | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "I've noticed that when I've been invited to view and comment on AI-safety related draft articles (in Google Docs), they tend to quickly attract a lot of extensive discussion, including from people who almost never participate on public forums like LessWrong or AI Alignment Forum. The number of comments is often an ord... |
45428735-5331-4500-9ffd-02ed1bcabc07 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Washington, D.C.: xkcd Discussion
Discussion article for the meetup : Washington, D.C.: xkcd Discussion
WHEN: 10 May 2015 03:00:00PM (-0400)
WHERE: Reynolds Center
We will be meeting in the Kogod Courtyard of the Donald W. Reynolds Center for American Art and Portraiture (8th and F Sts or 8th and G Sts NW,... |
8612b1de-045a-4d1b-b231-3a3c7e63250b | trentmkelly/LessWrong-43k | LessWrong | The Myth of the Myth of the Lone Genius
“Our species is the only creative species, and it has only one creative instrument, the individual mind and spirit of man. Nothing was ever created by two men. There are no good collaborations, whether in music, in art, in poetry, in mathematics, in philosophy. Once the miracle ... |
2a197ac1-36df-4767-a3cc-bdfd282ac430 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | [Link] PaLM 2 Technical Report
> We introduce PaLM 2, a new state-of-the-art language model that has better multilingual and reasoning capabilities and is more compute-efficient than its predecessor PaLM. PaLM 2 is a Transformer-based model trained using a mixture of objectives similar to UL2. Through extensive evalua... |
005df537-ddb8-40ff-8a00-db775786c8e2 | trentmkelly/LessWrong-43k | LessWrong | Chapter 4: What's the Problem?
To solve some task, you need to have it. What's the reason to have a task? Why do we work on it? Imagine someone has a job for you. Or you defined your mission by yourself. What do you trying to achieve?
I've described it like this: I am solving Problem.
But what is the Problem? When d... |
33fda6ce-9e8d-422f-8487-ea6e593879d3 | trentmkelly/LessWrong-43k | LessWrong | Optimization Amplifies
,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
I talk here about how a mathematician mindset can be useful for AI alignment. But first, a puzzle:
Given m, what is the least number n≥2 such that for 2≤k≤m, the base k representation of n consists entirely of 0s and 1s?
If you want to think about it yourself, ... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.