id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
fd68701c-4723-4a08-8fd5-5fae4c9e404c | trentmkelly/LessWrong-43k | LessWrong | Where do people discuss doing things with GPT-3?
I'm looking for examples of GPT-3 prompts to achieve certain effects. Is there a forum where this is discussed?
In particular, I'd like to see how people generate tweets. |
d60970a1-b917-4f56-a3d9-fbb765bccf57 | trentmkelly/LessWrong-43k | LessWrong | Instrumental Rationality: Postmortem
[In April, I set off to write a series of essays about instrumental rationality. Now that the project’s reached a pretty good stopping point, I’m looking back to see how my expectations and goals played out.]
Initial Goals:
[What I originally wanted the book to be like. My estima... |
45d167c2-e690-4f33-98b0-88190816f8c6 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Modeling AGI Safety Frameworks with Causal Influence Diagrams
We have written a paper that represents various frameworks for designing safe AGI (e.g., RL with reward modeling, CIRL, debate, etc.) as Causal Influence Diagrams (CIDs), to help us compare frameworks and better understand the corresponding agent incentives... |
1c61e290-5588-441e-9650-e480a652fa25 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Research agenda update
*“Our greatest fear should not be of failure, but of succeeding at something that doesn't really matter.” –DL Moody (allegedly)*
(Related: [Solving the whole AGI control problem, version 0.0001](https://www.lesswrong.com/posts/Gfw7JMdKirxeSPiAk/solving-the-whole-agi-control-problem-version-0-0... |
1ecc8a1d-4ba4-4ffc-88f9-afb461e13c39 | trentmkelly/LessWrong-43k | LessWrong | Simple Tricks to Improve KN95 Masks
There have already been a couple of posts on this topic, including this one which promotes a mask brace:
https://www.lesswrong.com/posts/CrikcGiaWK9CSSzrJ/i-made-an-n95-level-mask-at-home-and-you-can-too
Since then, there have been changes to the way we mask. Nowadays, I often see... |
875374f4-a4cc-440c-bdcc-e4570e471333 | trentmkelly/LessWrong-43k | LessWrong | Gender differences in spatial reasoning appear to be nurture
"In this study, we use a large-scale incentivized experiment with nearly 1,300 participants to show that the gender gap in spatial abilities, measured by time to solve a puzzle, disappears when we move from a patrilineal society to an adjoining matrilineal s... |
fdee755c-5336-4f5d-99a1-23ac1b1375be | trentmkelly/LessWrong-43k | LessWrong | What's the easiest way to currently generate images with machine learning?
As far as I understand, the images for the latest LessWrong books were generated by a neural net. I would like to generate images for the sequences I have on LessWrong in a similar way. What's the easiest way for me to do so? |
5b53e26f-c50e-4141-9646-8ba3a8402a45 | trentmkelly/LessWrong-43k | LessWrong | One-shot strategy games?
I'm looking for computer games that involve strategy, resource management, hidden information, and management of "value of information" (i.e. figuring out when to explore or exploit), which:
* *can* be beaten in 30 – 120 minutes on your first try (or, there's a clear milestone that's about t... |
b4b9ca6f-d8fa-4ee1-bf82-fc1f0243d839 | trentmkelly/LessWrong-43k | LessWrong | Updates and Reflections on Optimal Exercise after Nearly a Decade
Previously: https://www.lesswrong.com/posts/bZ2w99pEAeAbKnKqo/optimal-exercise
Firstly, do the basic epistemics hold up? As far as I know, yes. The basic idea that lifting twice a week and doing cardio twice a week add up to a calorie expenditure that ... |
7cc8cce9-f33b-48e1-a12d-3233a6d2ea25 | trentmkelly/LessWrong-43k | LessWrong | My thoughts on OpenAI's Alignment plan
This is my attempt at Eliezer's challenge.
https://www.lesswrong.com/posts/tD9zEiHfkvakpnNam/a-challenge-for-agi-organizations-and-a-challenge-for-1
My overall impression of this plan is that it is surprisingly good. Not that I am particularly surprised such a plan exists, if M... |
4036133a-e2e2-4acb-bc0f-840cfd27ce40 | trentmkelly/LessWrong-43k | LessWrong | Gelman Against Parsimony
In two posts, Bayesian stats guru Andrew Gelman argues against parsimony, though it seems to be favored 'round these parts, in particular Solomonoff Induction and BIC as imperfect formalizations of Occam's Razor.
Gelman says:
> I’ve never seen any good general justification for parsimony...
... |
61cfd609-d240-4744-9ac8-2570a8057eca | trentmkelly/LessWrong-43k | LessWrong | Stable agent, subagent-unstable
,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
Crossposted at the intelligent agent forum.
A reflectively consistent agent is one that is ok with creating copies with the same motivation as itself, or continuing to have the same motivation under self-modification.
A reflectively stable agent is o... |
3bb47f5e-a589-4c55-bab2-96145c491c5f | trentmkelly/LessWrong-43k | LessWrong | There is no royal road to alignment
Crossposted from the EA Forum: https://forum.effectivealtruism.org/posts/W4MfMmbwccDCRBfnj/there-is-no-royal-road-to-alignment
Epistemic status: tentative thoughts I've been processing for a while now and discussing mostly with philosophy folks.
Despite working in philosop... |
d288b218-60ee-405f-9c8e-5dad2d279ac3 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Melbourne LW Outing: Astronomy evening in Eltham, Saturday 29th June, 5:30pm
Discussion article for the meetup : Melbourne LW Outing: Astronomy evening in Eltham, Saturday 29th June, 5:30pm
WHEN: 29 June 2013 05:30:00PM (+1000)
WHERE: 54 Scenic Crescent Eltham North Victoria Australia
This event replaces t... |
a9251af6-7ecd-4e57-9125-fbfaaf3024de | trentmkelly/LessWrong-43k | LessWrong | Regularization Causes Modularity Causes Generalization
Epistemic Status: Exploratory
Things That Cause Modularity In Neural Networks
Modularity is when a neural network can be easily split into several modules: groups of neurons that connect strongly with each other, but have weaker connections to outside neurons. W... |
a5abe1d0-6c7f-403b-bebd-023ca87ee6f7 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Announcing the Harvard AI Safety Team
***TLDR**: The Harvard AI Safety Team (HAIST) is a new research-focused student group aimed at encouraging Harvard undergrads to pursue careers in technical AI safety. We think the club has been successful, and that other groups may have success following a similar format. We are ... |
1fb618bd-fa25-4c55-9c52-75f92fde14b8 | trentmkelly/LessWrong-43k | LessWrong | PrinciplesYou - Seems to be a new personality assessment tool
https://principlesyou.com
I saw a story that mentioned this tool and took the test. I probably didn't give the answers as much thought as perhaps I should have (but wouldn't that also fit with a good personality tool's analysis?) but found some of the resu... |
738b4332-6038-455a-9754-e230e6c4b4a5 | trentmkelly/LessWrong-43k | LessWrong | Learning is (Asymptotically) Computationally Inefficient, Choose Your Exponents Wisely
Epistemic status: A speculation based on personal observations, but looks coherent to me. (Sorry, no links or literature search)
Summary: Studying a subject gets progressively harder as you learn more and more, and the effort requi... |
7e785f74-1ac7-4d04-826b-ad05814707eb | trentmkelly/LessWrong-43k | LessWrong | Professing and Cheering
I once attended a panel on the topic, “Are science and religion compatible?” One of the women on the panel, a pagan, held forth interminably upon how she believed that the Earth had been created when a giant primordial cow was born into the primordial abyss, who licked a primordial god into exi... |
dc106a49-9ef6-446d-b830-140ebfd47f03 | trentmkelly/LessWrong-43k | LessWrong | Interpreting Complexity
This is a cross-post - as some plots are meant to be viewed larger than LW will render them (and on a dark background) it is recommended this post be read via the original site.
Thanks to Zach Furman for discussion and ideas and to Daniel Murfet, Dmitry Vaintrob, and Jesse Hoogland for feedba... |
11e9678a-4d16-4b5d-a09a-7e589f5e0d85 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Understanding the Capabilities, Limitations, and Societal Impact of Large Language Models
Introduction
------------
On October 14th, 2020, researchers from OpenAI, the Stanford Institute for Human-Centered Artificial Intelligence, and other universities convened to discuss open research questions surrounding [GPT-3... |
9691da03-653c-4a66-aa1d-42647f7748e6 | trentmkelly/LessWrong-43k | LessWrong | Tormenting Gemini 2.5 with the [[[]]][][[]] Puzzle
Includes a link to the full Gemini conversation, but setting the stage first:
There is a puzzle. If you are still on Tumblr, or live in Berkeley where I can and have inflicted it on you in person[1], you may have seen it. There's a series of arcane runes, composed e... |
4498f728-50c4-4716-89c8-06d471c0cc1f | trentmkelly/LessWrong-43k | LessWrong | Deleted comments archive?
This is a question for the Less Wrong moderation team.
Suppose that I post a comment on a post, and the author of that post deletes that comment. Now, that comment isn’t lost, per se (as its contents are sent to me via private message), but it’s no longer publicly accessible. This is likely ... |
b00fe4ec-6d7e-40df-8276-1ae674d900f5 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | How Much Computational Power Does It Take to Match the Human Brain?
Joe Carlsmith with a really detailed report on computational upper bounds and lower bounds on simulating a human brain:
> Open Philanthropy is interested in when AI systems will be able to perform [various tasks](https://www.openphilanthropy.org/bl... |
7cc42106-ce94-45b3-bf4b-ea82159e0597 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Analysis of AI Safety surveys for field-building insights
Introduction
============
To help grow the pipeline of AI safety researchers, I conducted a project to determine how demographic information (e.g. level of experience, exposure to AI arguments) affects AI researchers’ responses to AI safety. In addition, I exa... |
b9931886-abc6-4b8d-9768-f9bcc427a27f | trentmkelly/LessWrong-43k | LessWrong | An example elevator pitch for AI doom
I have been surprised to repeatedly see the claim that there isn't even an argument for concern about AI. That the claim has been made without evidence and can therefore be dismissed.
Obviously, there is an extensive library of evidence and arguments that have been made for deca... |
88ca76e3-4ce3-44a0-8c4a-cb877e32933d | trentmkelly/LessWrong-43k | LessWrong | Mystical science
Recently I heard an author interviewed on NPR about his book. I can no longer remember the book's name or topic; but I remember that a couple of times during the interview, the author made puzzling categorizations. One of them was approximately: "Throughout history, there are two forces that act on... |
08b6081b-7a10-48dd-9bfa-59930c272425 | trentmkelly/LessWrong-43k | LessWrong | Open-ended/Phenomenal Ethics (TLDR)
This is a short version of the more complete post
More clarity and context here
I formulate a desiderata/procedure (called "phenomenal ethics" or "open-ended ethics") :
Expanding the action space and autonomy of a maximum of phenomenons (enhancing ecosystemic values, and t... |
fb0a650b-a204-416d-a5a8-24dc8089be26 | trentmkelly/LessWrong-43k | LessWrong | Extended Interview with Zhukeepa on Religion
Introduction from Ben
Zhukeepa is a LessWronger who I respect and whose views I'm interested in. In 2018 he wrote the first broadly successful explication of Paul Christiano's research ideas for AI alignment, has spent a lot of time interviewing people in AI about their per... |
6a64b5e0-1b41-4327-8817-85e284c489d3 | trentmkelly/LessWrong-43k | LessWrong | Help! What do I say at my wedding?
Update: This is what I wrote as my vow
> Dearly beloved, believe me when I say that, back in 2019, I would have gladly dug myself a nuke shelter to avoid your vibrant radiance.
>
> Yet a few days ago, late at night and settling into the quiet comfort of your soothing voice, I thoug... |
d52e69fb-72af-4b72-b770-5e02f7c0fb59 | trentmkelly/LessWrong-43k | LessWrong | Long-Term Future Fund: April 2023 grant recommendations
(Cross-posted from the EA Forum.)
Introduction
This payout report is meant to cover the Long-Term Future Fund's grantmaking starting January 2022 (after our December 2021 payout report), going through April 2023 (1 January 2022 - 30 April 2023).
* Total fundi... |
a3423e29-e914-47f6-aba8-5c71153c1c2d | trentmkelly/LessWrong-43k | LessWrong | Science - Idealistic Versus Signaling
[This is a version of an first draft essay I wrote for my blog. I intend to write another version, but it is going to take some time to research, and I want to get this out where I can start getting some feedback and sources for further research.]
The responses to the recent lea... |
74075482-fd5f-4e0c-85d3-644234d46074 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Bratislava
Discussion article for the meetup : Bratislava
WHEN: 10 November 2014 06:00:00PM (+0100)
WHERE: Bistro The Peach, Mariánska 3, Bratislava
Bistro the Peach asi stále nie je poriadne označené, ale je to oproti predajni richmanov na Mariánskej. Alebo sa riaďte podľa google mapy a čísla ulice. Vchod... |
03344e79-d193-4b6a-a6e0-e893615c2283 | trentmkelly/LessWrong-43k | LessWrong | Weekly LW Meetups
This summary was posted to LW Main on November 21st. The following week's summary is here.
New meetups (or meetups with a hiatus of more than a year) are happening in:
* Bangalore Meetup: 22 November 2014 04:10PM
Irregularly scheduled Less Wrong meetups are taking place in:
* [Atlanta] MIRIxAtl... |
c8765f6d-bc69-4d00-8088-7e130f5756e5 | trentmkelly/LessWrong-43k | LessWrong | Designing serious games - a request for help
We need some ideas for serious games. Games that will help us be better. Games that reward us for improving ourselves (even if just by the satisfaction of seeing our scores improve). Games that will help us in our quest of Tsoyoku Naritai
We've got an upcoming hackday in L... |
b12cc4e2-d4c4-4b5a-b803-44f6d2f399c5 | trentmkelly/LessWrong-43k | LessWrong | Lessons on AI Takeover from the conquistadors
(Talk given at an event on Sunday 28th of June. Daniel Kokotajlo is responsible for the talk, David Lambert edited the transcript.
If you're a curated author and interested in giving a 5-min talk, which will then be transcribed and edited, sign up here.)
Daniel Koko... |
c2cbad7a-1e3d-4522-ae0e-c643329ef2a4 | StampyAI/alignment-research-dataset/blogs | Blogs | September 2021 Newsletter
Scott Garrabrant has concluded the main section of his [Finite Factored Sets](https://www.lesswrong.com/s/kxs3eeEti9ouwWFzr) sequence (“Details and Proofs”) with posts on [inferring time](https://www.lesswrong.com/s/kxs3eeEti9ouwWFzr/p/hePucCfKyiRHECz3e) and [applications, future work, and sp... |
a1a856fe-3c07-47a2-a5d6-851f62ab0226 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Formalizing Two Problems of Realistic World Models
I'm pleased to announce a new paper from MIRI: [Formalizing Two Problems of Realistic World Models](https://intelligence.org/files/RealisticWorldModels.pdf).
Abstract:
>
> An intelligent agent embedded within the real world must reason about an environment which... |
fcc167b9-8368-4d88-ada8-722131dd9c1a | trentmkelly/LessWrong-43k | LessWrong | Meetup : Washington DC Effective Altruism Meetup
Discussion article for the meetup : Washington DC Effective Altruism Meetup
WHEN: 12 January 2014 03:00:00PM (-0500)
WHERE: National Portrait Gallery, Washington, DC 20001, USA
We'll be meeting to discuss effective altruism, or whatever else people want to talk about... |
6b44a50d-3dcf-46d3-990a-034d0f74a006 | trentmkelly/LessWrong-43k | LessWrong | [Stanford Daily] Table Talk
This is my favorite article in Stanford's student newspaper: the Stanford Daily. It raises the question: why are students at an institution so prestigious... so boring?
At this late hour of Stanford history, can anyone hope to bring up a serious subject in a dining hall? Every so often I l... |
c52e29a2-0989-450b-a2de-e6738b4948f0 | trentmkelly/LessWrong-43k | LessWrong | "Can We Survive Technology" by von Neumann
> "The great globe itself" is in a rapidly maturing crisis —a crisis attributable to the fact that the environment in which technological progress must occur has become both undersized and underorganized. To define the crisis with any accuracy, and to explore possibilities of... |
830a5dad-e03f-442c-985f-e2c2978f6a3d | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Efficient Cross-Domain Optimization
Today's post, Efficient Cross-Domain Optimization was originally published on 28 October 2008 . A summary (taken from the LW wiki):
> To speak of intelligence, rather than optimization power, we need to divide optimization power by the resources needed, or the amount... |
7c87aade-8669-4955-8d5d-e1d65c9b2a18 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Top lesson from GPT: we will probably destroy humanity "for the lulz" as soon as we are able.
Forget complicated "sharp left turn" schemes, nefarious nanobots, lists of lethalities, out-of-distribution actions, failed AI boxing. As Zvi pointed out in multiple posts, like [this one](https://thezvi.wordpress.com/2023/04... |
10132fb9-f900-4e5c-bce4-81cbf9fdf576 | trentmkelly/LessWrong-43k | LessWrong | AlphaGo versus Lee Sedol
There have been a couple of brief discussions of this in the Open Thread, but it seems likely to generate more so here's a place for it.
The original paper in Nature about AlphaGo.
Google Asia Pacific blog, where results will be posted. DeepMind's YouTube channel, where the games are being l... |
ac82c587-3da5-4701-9075-213a38dab88e | trentmkelly/LessWrong-43k | LessWrong | Understanding Gradient Hacking
This post was written under Evan Hubinger’s direct guidance and mentorship, as a part of the Stanford Existential Risks Institute ML Alignment Theory Scholars (MATS) program.
Additional thanks to Michele Campolo, Oliver Zhang, Adam Shimi, and Leo Gao for their thoughts and feedback on t... |
789e1c86-cb34-4e70-8f99-eed5e0ed00f2 | trentmkelly/LessWrong-43k | LessWrong | Concrete examples of doing agentic things?
I've been thinking about agency recently. I want to be more agentic!
Sometime last year, I read Cultivating And Destroying Agency and Seven ways to become unstoppably agentic. I recently saw How to be More Agentic and it kind of spurred me to write this post (which is actual... |
aaf51b9a-0a18-402f-81f0-7bb7dbc20229 | trentmkelly/LessWrong-43k | LessWrong | Quantified Intuitions: An epistemics training website including a new EA-themed calibration app
Crossposted to EA Forum
TL;DR Quantified Intuitions helps users practice assigning credences to outcomes with a quick feedback loop. Please leave feedback in the comments, join our Discord, or send thoughts to aaron@sage-f... |
d7d54a4a-c250-4e0d-919a-a7cf406f8469 | trentmkelly/LessWrong-43k | LessWrong | A simple web app to create aversion to unhealthy food
At a startup camp weekend, I've made this (incredibly sketchy, amateurish) app (NSFW) which implements classical conditioning to create a negative response to candy, smoking, or meat.
After using it intensively for about a day while writing it, I certainly feel di... |
fa9640cc-33ec-4407-8230-f7312fb366b0 | trentmkelly/LessWrong-43k | LessWrong | Untapped Potential at 13-18
The Western world is facing a crisis of stretched adolescence, keeping children exclusively "learning" for longer than necessary.
Imagine that you are a student, around 14 years old. You fall into the top 5%, not the extremely motivated startup founder or math prodigy that you hear about ... |
923ddd64-cf55-44cb-82ec-25d9a207c760 | trentmkelly/LessWrong-43k | LessWrong | Framery Phone Booth CO2 Accumulation
The lab the NAO works out of has a Framery O phone booth, for taking calls. It's very good at sound isolation, but being so small (~1.5 m3, 54 ft3) you might be worried about CO2 building up:
I decided to test it. Here are two rounds of sitting in the booth until the meter (Temp... |
a839c030-b813-48ee-93f1-d22805dbd6e2 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | My research methodology
(*Thanks to Ajeya Cotra, Nick Beckstead, and Jared Kaplan for helpful comments on a draft of this post*.)
I really don’t want my AI to strategically deceive me and resist my attempts to correct its behavior. Let’s call an AI that does so ***egregiously misaligned*** (for the purpose of this po... |
621af655-fe4f-4b8b-9c18-c0c8ec97de6a | trentmkelly/LessWrong-43k | LessWrong | Entropy Scaling And Intrinsic Memory
This is the second post in a series that explores a circle of ideas around entropy and information processing in complex systems. We hope it will form an accessible introduction to James Crutchfield's Computational Mechanics framework.
In the previous post we saw that
* Data ... |
cb590143-90b1-476a-81e0-71b2075a67aa | trentmkelly/LessWrong-43k | LessWrong | What's next for instrumental rationality?
Preceded by: Curating "The Epistemic Sequences" (list v.0.1)
Epistemic status: speculations and ideations from me about the potential for further progress on broadly accessible instrumental rationality content.
In Curating "The Epistemic Sequences", I explained how the epist... |
5dd63516-df82-4e7c-9641-9113c0ce6bcb | awestover/filtering-for-misalignment | Redwood Research: Alek's Filtering Results | id: post3839
I recommend reading Scalable AI control before reading this post. In particular, in the "Hard values, easy values" section of this post, Paul writes: My statement of the control problem is only really meaningful because there are instrumental subgoals that are shared (or are extremely similar) between man... |
2b723432-20c6-468d-96ab-88a6666cb7dd | StampyAI/alignment-research-dataset/lesswrong | LessWrong | What are some examples of AIs instantiating the 'nearest unblocked strategy problem'?
A paragraph explaining the problem, from Ngo, Chan, and Mindermann ([2023](https://arxiv.org/abs/2209.00626)). I've bolded the key part:
> Our definition of internally-represented goals is consistent with policies learning multiple... |
827f4174-5837-43cc-941b-67f857e508d2 | trentmkelly/LessWrong-43k | LessWrong | Want to help me test my Anki deck creation skills?
I'm interested in trying to make better Anki decks for the LessWrong community, but I want to see how well I can actually do this first. There's a lot of knowledge out there about how to format and create decks, but it's still a decent amount of work, and there are lo... |
c6a6314e-e6c4-4fc3-810f-089a6a6cf65f | trentmkelly/LessWrong-43k | LessWrong | October gwern.net links
None |
78c28954-e416-4def-89bd-364d3a6e1aae | StampyAI/alignment-research-dataset/lesswrong | LessWrong | A great talk for AI noobs (according to an AI noob)
I'm far from an expert on AI and alignment but I really liked this [Ted talk](https://forum.effectivealtruism.org/out?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DC_78DM8fG6E) and think it serves as a good introduction to the topic. Here are some highlights:
It s... |
fce97349-2ff6-494d-981a-1ab12c841f92 | trentmkelly/LessWrong-43k | LessWrong | You can use GPT-4 to create prompt injections against GPT-4
So, I saw this tweet. It turns out you can ask GPT4 to compress a message in a way it can understand. The end result is usually not understandable for a human. It's interesting, since it allows you to effectively rise the context length. It also reveals a spo... |
497156b5-94e5-4b50-9fcf-b80732ba8891 | trentmkelly/LessWrong-43k | LessWrong | AI misalignment risk from GPT-like systems?
Right now, it seems that the most likely way we're gonna get an (intellectually) universal AI is by scaling models such as GPT. That is, models trained by self-supervised learning on massive piles of data, perhaps with a similar architecture to the transformer.
I do not se... |
72a459d7-4878-4601-af3e-291e3ef6ce85 | trentmkelly/LessWrong-43k | LessWrong | AI safety raising awareness resources bleg
> Bleg definition
>
> (Internet slang) An entry on a blog requesting information or contributions. I posted a bleg in the hope of learning more about local tourism.
>
> https://www.yourdictionary.com/bleg
Most ordinary people don't know about AI safety. They might have hea... |
bdfcfe6b-2fee-48d3-bc88-55c85153e68a | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "Since the episode with Eugine_Nier, I have received three private messages from different people asking me to investigate various cases of suspected mass downvoting. And to be quite honest, I don't want to deal with this. Eugine's case was relatively clear-cut, since he had engaged in systematic downvoting of a massiv... |
c2bb8c08-7d1b-4ddb-b9af-676ff5cddb55 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Mimicking Behaviors in Separated Domains
1 Introduction
---------------
*Mimicking* a behavior from a system A𝐴Aitalic\_A to a system B𝐵Bitalic\_B is a
common practice in Computer Science (CS) and Software Engineering (SE).
Examples include a robot that has to real-time adapt a human behavior (?),
or simultaneous... |
2444ed15-856b-4840-922c-0f7bf46e72a7 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Consequentialists: One-Way Pattern Traps
*Generated during MATS 2.1.*
*A distillation of my understanding of Eliezer-consequentialism.*
*Thanks to Jeremy Gillen, Ben Goodman, Paul Colognese, Daniel Kokotajlo, Scott Viteri, Peter Barnett, Garrett Baker, and Olivia Jimenez for discussion and/or feedback; to Eliezer Yu... |
d40a84da-41c4-4596-9294-364f38370e84 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | VLM-RM: Specifying Rewards with Natural Language
tl;dr
-----
We show how to use Vision-Language Models (VLM), and specifically CLIP models, as reward models (RM) for RL agents. Instead of manually specifying a reward function, we only need to provide text prompts like “a humanoid robot kneeling” to instruct and provi... |
21c65235-08b0-4406-8bed-a6ccf44dfa37 | trentmkelly/LessWrong-43k | LessWrong | Lecture Series on Tiling Agents
For my AISC, I'll[1] be presenting more details about the research every Thursday for approximately the next three months. If you are interested in listening in, here is a calendar link.
EDIT:
The calendar link apparently doesn't invite people to the recurring event; I'm not sure I ca... |
e6300982-fbc3-454a-99a7-aced1827db4c | trentmkelly/LessWrong-43k | LessWrong | Emotional installation of concepts
This is a plausible theory I've got-- I don't know if formal work has been done in the area.
I used to read discussions of the definition of science fiction until I came to the conclusion that there was no point. People have strongly felt intuitions about what science fiction really... |
c3697a24-a535-4309-b1a4-6b30dfb7e8ae | trentmkelly/LessWrong-43k | LessWrong | UDT can learn anthropic probabilities
Today I realized that UDT doesn't have to clash with "objective" views of anthropic probability, like SSA or SIA: instead it can, in some sense, learn which of these views is true!
The argument goes like this. First I'll describe a game where SSA and SIA lead to different decisio... |
7106aafb-39ee-438c-b7f9-91d40ab74287 | trentmkelly/LessWrong-43k | LessWrong | Best LessWrong posts for understanding scientific studies.
For someone who is not totally quantitatively literate, or whose background is outside of STEM, what are the best resources for understanding scientific studies? I'm looking for books, articles, or posts that explain both the different parts of these studies a... |
21f46ca4-e899-473e-bbf5-fc74f60bb6c6 | trentmkelly/LessWrong-43k | LessWrong | Meetup : West LA Meetup - Quantifying and Measuring for Self-Experimentation
Discussion article for the meetup : West LA Meetup - Quantifying and Measuring for Self-Experimentation
WHEN: 05 June 2013 11:00:53PM (-0700)
WHERE: 10850 West Pico Blvd, Los Angeles, CA 90064
When: 7:00pm Wednesday, June 5th.
Where: The ... |
2186f514-0771-40d2-ba65-cd1bd9943c42 | trentmkelly/LessWrong-43k | LessWrong | A culture of exploitation?
Thoughts in the car:
I just had a discussion with a coworker who bemoans the lack of respect and responsibility of the newer generations. While I do think these concerns are typically overstated due to availability bias, there could be some truth to them.
I hypothesize that a significant c... |
37cc1f1a-092d-4802-b8b7-e2715aa32e57 | trentmkelly/LessWrong-43k | LessWrong | ETH Zurich, ML related MSc programs
This article is part of a series of articles on different European master's programs related to artificial intelligence and machine learning.
The purpose of this post is to give an overview of Machine Learning related Master’s programs at ETH Zurich. Machine Learning is a topic tha... |
22b98da7-104e-4e01-baac-8fb2d270ce08 | trentmkelly/LessWrong-43k | LessWrong | Predictive coding & depression
Epistemic status: wild speculation.
This is a follow-up to the 2017 blog post Slate Star Codex: Toward a Predictive Theory of Depression. I think that post has fundamentally the right idea, but that it had some pieces of the puzzle that didn't quite fit satisfyingly.
Well I personally ... |
7d7516c3-f021-49bd-b9b1-43fdfbc04932 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | Responses to apparent rationalist confusions about game / decision theory
I’ve encountered various claims about how AIs would approach game theory and decision theory that seem pretty importantly mistaken. Some of these confusions probably aren’t that big a deal on their own, and I’m definitely not the first to point ... |
1ea4eb81-082d-4f7d-8c53-f300597eca61 | trentmkelly/LessWrong-43k | LessWrong | Book Club Update, Chapter 3 of Probability Theory
Previously: Book Club introductory post - Chapter 1 - Chapter 2
We will shortly move on to Chapter 3 (I have to post this today owing to vacation - see below). I have updated the previous post with a summary of chapter 2, with links to the discussion as appropriate. B... |
fb685fe6-b2ec-40e6-9ac2-af6e30246f24 | trentmkelly/LessWrong-43k | LessWrong | More Stubs of Posts I want to Write
Followup to prior stubs of posts I want to write. Progress:
* Archipelago got written.
* Recognizing Demon Threads is on its way.
* I feel like someone wrote something about Social Reality but I'm not sure.
* I wrote things about the personal standards I wanted to cultivate but... |
25a03a3f-308b-4452-828e-dcc96caeab6f | trentmkelly/LessWrong-43k | LessWrong | Scott Adams: 'Mockability Test'
I'm not sure what the protocol is for linking to or quoting another site on LW, but I thought this would fit here for two reasons: first, because I'm curious what people here think about his 'mockability' test, which seems to be half in jest but (I think) has a serious point at its core... |
b46dd9f5-5146-4a71-91a6-50540d2ec6cd | trentmkelly/LessWrong-43k | LessWrong | Ways to buy time
In our last post, we claimed:
> On the margin, we think more alignment researchers should work on “buying time” interventions instead of technical alignment research (or whatever else they were doing).
But what does “buying time” actually look like? In this post, we list some interventions that have... |
7f917ce3-3969-4680-bd1b-2d94e738c2cb | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | 5. Risks from preventing legitimate value change (value collapse)
The second risk that follows from the VCP concerns the obstruction of legitimate value change. In what follows, I will consider the mechanisms by which legitimate value change might come to be impeded, and whether and when this might constitute an activ... |
7baa91c6-c273-449d-a657-5b2c29a6dc00 | trentmkelly/LessWrong-43k | LessWrong | Elon Musk announces xAI
Some quotes & few personal opinions:
FT reports
> Musk is also in discussions with a number of investors in SpaceX and Tesla about putting money into his new venture, said a person with direct knowledge of the talks. “A bunch of people are investing in it . . . it’s real and they are excited ... |
47af9e9b-9361-4556-969c-32936792bc40 | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] Of Gender and Rationality
Today's post, Of Gender and Rationality was originally published on 16 April 2009. A summary (taken from the LW wiki):
> Analysis of the gender imbalance that appears in "rationalist" communities, suggesting nine possible causes of the effect, and possible corresponding soluti... |
949bdbae-6761-46d9-93e6-ae87828dfa2c | trentmkelly/LessWrong-43k | LessWrong | The Wrong Side of Risk
Part 4 of the Inefficient Markets sequence.
[Formerly entitled How to Pick the Wrong Side, which on reflection, I felt was too clickbaity. I am not your financial advisor! For educational purposes only. Read part 1 first.]
The Basic Idea of Insurance
You want protection from a risk. An insure... |
927852f2-f322-4422-b119-a11486bde6be | trentmkelly/LessWrong-43k | LessWrong | [SEQ RERUN] The Cartoon Guide to Lob's Theorem
Today's post, The Cartoon Guide to Löb's Theorem was originally published on 17 August 2008. A summary (taken from the LW wiki):
> An explanation, using cartoons, of Lob's theorem.
Discuss the post here (rather than in the comments to the original post).
This post ... |
37d39e40-e648-4132-a731-bf2f9c60e3e2 | trentmkelly/LessWrong-43k | LessWrong | How a pathological procrastinor can lose weight [Anti-akrasia]
[This post has now been subsumed by the following: blog.beeminder.com/akrasia. Also, the service described below, then known as Kibotzer, is now a real startup called Beeminder, announced here: http://lesswrong.com/lw/7z1/antiakrasia_tool_like_stickkcom_fo... |
9908a2ea-3e06-4f01-9abe-60bea3aedc40 | trentmkelly/LessWrong-43k | LessWrong | Voting Weight Discussion
One of the distinguishing features of LW 2.0 is Voting Weight. People with more karma will be able to give higher-weighted upvotes and downvotes.
This is part of an overall plan to make it better at sorting signal-from noise. But it's a major change to the system and seemed worthy of a dedica... |
91905710-02ce-4e17-a818-f1456c96d989 | trentmkelly/LessWrong-43k | LessWrong | Thiel on AI & Racing with China
This post is a transcript of part of a podcast with Peter Thiel, touching on topics of AI, China, extinction, Effective Altruists, and apocalyptic narratives, published on August 16th 2024.
If you're interested in reading the quotes, just skip straight to them, the introduction is not ... |
3e30702d-af56-4299-ba7c-2c1ac8787bf7 | trentmkelly/LessWrong-43k | LessWrong | Weekly LW Meetups
This summary was posted to LW Main on December 27th. The following week's summary is here.
Other irregularly scheduled Less Wrong meetups are taking place in:
* Montreal - How to Actually Change your Mind: 07 January 2014 07:00PM
The remaining meetups take place in cities with regular scheduling,... |
1dd4b1ad-9359-45ab-bc5b-eafb1ffcae6c | trentmkelly/LessWrong-43k | LessWrong | Durkon, an open-source tool for Inherently Interpretable Modelling
or, what abstractapplic has been tinkering with for the past two years, and how it can be applied to fictional card games
Motivation
Interpretability of ML models is important. If people don’t understand how a model makes its predictions, they will b... |
0b1fbb9c-43bf-4e27-9b8a-12cdf2dede83 | trentmkelly/LessWrong-43k | LessWrong | Every niche event should also be a meetup
I was just at a Tyler Cowen lecture in NYC — one of my heroes and biggest intellectual influences — and watching everyone file out after the talk, I kept thinking about this really great comment someone had made on the event announcement: “Do these 92Y talks typically offer sp... |
9645f7c4-3316-4b00-a6d6-b48096b6a14d | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Why Subagents?
The justification for modelling real-world systems as “agents” - i.e. choosing actions to maximize some utility function - usually rests on [various coherence theorems](https://www.lesswrong.com/posts/RQpNHSiWaXTvDxt6R/coherent-decisions-imply-consistent-utilities). They say things like “either the syst... |
04c33200-1528-4796-91c7-4ec5f137c3d9 | trentmkelly/LessWrong-43k | LessWrong | General-purpose forecasting and the associated community
This post includes some lists gathered by web surfing and link-traipsing, as well as some broad qualitative impressions. The qualitative impressions are not always thoroughly grounded, and some of them may be plain wrong.
As contract work for MIRI, I've been re... |
2dbb1b8f-143a-410e-b171-06740704b8f5 | trentmkelly/LessWrong-43k | LessWrong | When is it Wrong to Click on a Cow?
[Speculative philosophy. Originally on Grand, Unified, Crazy.]
Three Stories
Imagine, for a moment, three young adults recently embarked on the same promising career path. The first comes home from work each day, and spends their evenings practising and playing a musical instrumen... |
39008a38-d2dd-474c-947d-1ff9ef4ef671 | trentmkelly/LessWrong-43k | LessWrong | [linkpost] "What Are Reasonable AI Fears?" by Robin Hanson, 2023-04-23
Selected quotes (all emphasis mine):
> Why are we so willing to “other” AIs? Part of it is probably prejudice: some recoil from the very idea of a metal mind. We have, after all, long speculated about possible future conflicts with robots. But par... |
e100b5c1-c77a-4d49-9d96-fc8839250897 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Evan Hubinger on Homogeneity in Takeoff Speeds, Learned Optimization and Interpretability
Below is the transcript of my chat with Evan Hubinger, interviewed in the context of [the inside view](https://anchor.fm/inside-view) a podcast about AI Alignment. The links below will redirect you to corresponding timestamps in ... |
c3e44914-42df-4f87-8e2e-fcd1de0a9b4c | trentmkelly/LessWrong-43k | LessWrong | Meetup : Warsaw February Meetup
Discussion article for the meetup : Warsaw February Meetup
WHEN: 21 February 2015 06:00:00PM (+0100)
WHERE: Warsaw
The next meetup in Warsaw - we are slowly getting regular!
Facebook event: https://www.facebook.com/events/436771293136918 Facebook group for better coordination: https... |
9b33a842-f384-4238-b9bf-c2e51970f229 | trentmkelly/LessWrong-43k | LessWrong | [AN #133]: Building machines that can cooperate (with humans, institutions, or other machines)
Alignment Newsletter is a weekly publication with recent content relevant to AI alignment around the world. Find all Alignment Newsletter resources here. In particular, you can look through this spreadsheet of all summaries ... |
0fc07af1-2c6a-422e-a9e9-91ae90ed87ea | trentmkelly/LessWrong-43k | LessWrong | Normative assumptions: answers, emotions, and narratives
Agents can be modelled as pairs (p, R), where R is a reward, and p is a planner that takes R and more or less rationally outputs the policy p(R).
Normative assumptions are assumptions that can distinguish between different pairs (p, R) and (p', R'), despite tho... |
c6ccc9dd-d06f-4ca0-b999-0e8286eb07d0 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | ea.domains - Domains Free to a Good Home
**Tl;dr: I’ve set up**[**a database of domains at ea.domains**](https://ea.domains/) **which are**[**free to a good home**](https://ea.domains/#about)**, to prevent them being squatted and blocked for use. You can add domains you control to it using**[**this form**](https://cod... |
80d09501-ba68-4dc3-9b29-bb501d8900cb | trentmkelly/LessWrong-43k | LessWrong | Don't Use the "God's-Eye View" in Anthropic Problems.
This is another attempt to promote my solution to the anthropic paradoxes: Perspective-Based Reasoning. Partially motivated by Joe Carlsmith's series on SSA vs SIA.
SSA, SIA, and PBR
They are many ways to state SSA and SIA with equivalent effects on calculati... |
390d404c-6b8d-43f3-accf-726fdf0231c7 | trentmkelly/LessWrong-43k | LessWrong | Meetup advice sequence intro
I'm transitioning into working on global meetup organization as more of a full-time thing, and the lowest-hanging fruit — something many people have asked for — seemed to be writing up all the advice I have on how to run a meetup well. So that's my first project!
Over the past few years I... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.