id stringlengths 36 36 | source stringclasses 15
values | formatted_source stringclasses 13
values | text stringlengths 2 7.55M |
|---|---|---|---|
c1735725-6e1a-44eb-852e-4e9804a5ec28 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Universal Adversarial Triggers for Attacking and Analyzing NLP WARNING: This paper contains model outputs which are offensive in nature.
1 Introduction
---------------
| | | |
| --- | --- | --- |
| Task | Input ( red = trigger) | Model Prediction |
|
Sentiment
Analysis
| zoning tapping fiennes Visual... |
04f80832-5062-4156-9988-d4c05726107a | StampyAI/alignment-research-dataset/arbital | Arbital | Colon-to notation
In mathematics, the notation $f : X \to Y$ (here, "colon-to notation," because the arrow $\to$ is written "\to" in [https://arbital.com/p/LaTeX](https://arbital.com/p/LaTeX)) means that $f$ is a [function](https://arbital.com/p/3jy) with [domain](https://arbital.com/p/3js) $X$ and [codomain](https://... |
2b54d8e4-e752-474b-8720-4da118d110f4 | trentmkelly/LessWrong-43k | LessWrong | What's the most annoying part of your life/job?
Hi, I'm an entrepreneur looking for a startup idea.
In my experience, the reason most startups fail is because they never actually solve anyone's problem. So I'm cheating and starting out by identifying a specific person with a specific problem.
So I'm asking you, what... |
b5099ac9-abc7-4970-8448-aa186d62a513 | trentmkelly/LessWrong-43k | LessWrong | An Analogy for Interpretability
I.
Imagine newly baked interpretability researcher Namor, who claims that he solved interpretability. He says: “This is a bunch of matrices with a bunch of floating point values which are then multiplied, what’s the big deal?” His point of view might be considered myopic, but I think it... |
5fcd2f93-fefc-4da3-9b0a-c8366bcb6b28 | trentmkelly/LessWrong-43k | LessWrong | Green Line Extension Opening Dates
Talk of extending the Green Line from Lechmere to Ball Square and Tufts goes back a very long way, but it's finally open!
In celebration, I wanted to go back and collect the completion dates we've been given over the years:
* 1990-12-19: 2010
* 1994-10-04: 2011-12-31
* 200... |
29275e47-5b06-4515-8b0b-d8bca21f34ce | trentmkelly/LessWrong-43k | LessWrong | Reply to Duncan Sabien on Strawmanning
> DAVID: [...] Because that's all we can ever have of each other: an imitation, a Copy. All we can ever know about are the portraits of each other inside our own skulls.
> LORAINE: Is that all you think I am? An idea in your head?
> DAVID: No! But if it's all I have, then it's al... |
850219ee-1de9-4b46-9864-39e0cdadbfd9 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | What considerations influence whether I have more influence over short or long timelines?
As my timelines have been shortening, I've been rethinking my priorities. As have many of my colleagues. It occurs to us that there are probably general considerations that should cause us to weight towards short-timelines plans,... |
cf7f0444-7db5-4f37-83f5-6263242a9b82 | trentmkelly/LessWrong-43k | LessWrong | How to Contribute to the Coronavirus Response on LessWrong
The LessWrong team is currently focusing full-time on the coronavirus (Why?). We’ve made some pretty big asks with open questions like The LessWrong Research Agenda, and we’re super appreciative of everyone who’s participated and generally written really usefu... |
649c66f1-7dc7-4cec-9a81-ab687e3acbab | trentmkelly/LessWrong-43k | LessWrong | Good Heart Donation Lottery
Note: This was for April Fools. I'll leave the original top text for posterity, but see the updates for the current state of actual donation lottery.
Greetings, all!
Good Heart Tokens are clearly the mechanism we need to do more Good in the world. To that end, I'm starting a donation lott... |
a445e4d2-ef34-413b-98de-6fe351f56b34 | trentmkelly/LessWrong-43k | LessWrong | Guided By The Beauty Of Our Weapons
[Content note: kind of talking around Trump supporters and similar groups as if they’re not there.]
I.
Tim Harford writes The Problem With Facts, which uses Brexit and Trump as jumping-off points to argue that people are mostly impervious to facts and resistant to logic:
> All th... |
48da880b-174f-47ec-a2f7-0b74f295918e | trentmkelly/LessWrong-43k | LessWrong | Playlist Inspired by Manifest 2024
Okay, I think it's time to stop polishing this playlist & post it. This is music about the moods of people I met at Manifest. I'm pretty delighted with it! |
c73ae5ec-2605-4d38-9433-ed9824c8ccf1 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | What new psychology research could best promote AI safety & alignment research?
I've got a sabbatical coming up for 2024, and, as a psych professor concerned about AI X risk, I'd like to spend the year doing some empirical psychology research that helps inform AI safety and alignment.
What are some key questions on ... |
80e5a218-f319-43fb-9dfa-e661e99a248a | trentmkelly/LessWrong-43k | LessWrong | Public thread for researchers seeking existential risk consultation
LW is one of the few informal places which take existential risk seriously. Researchers can post here to describe proposed or ongoing research projects, seeking consultation on possible X-risk consequences of their work. Commenters should write thei... |
dbc18de3-4819-4234-88db-8c657af76278 | trentmkelly/LessWrong-43k | LessWrong | Built-in Measuring Spoons
In making the kids breakfast each morning, which has moved on from crepes and waffles to pancakes, I want a consistent amount of baking powder. I realized that the container has a built-in "measuring spoon". If you shake it with the lid on, powder gets up above the leveling shelf:
Take the... |
25194b33-6ab5-4297-bd57-2f9f79c9e3f4 | StampyAI/alignment-research-dataset/blogs | Blogs | Shulman and Yudkowsky on AI progress
This post is a transcript of a discussion between Carl Shulman and Eliezer Yudkowsky, following up on [a conversation with Paul Christiano and Ajeya Cotra](https://www.lesswrong.com/posts/7MCqRnZzvszsxgtJi/christiano-cotra-shulman-and-yudkowsky-on-ai-progress).
Color key:
... |
5f019fa1-fe38-420f-8e05-b044f23faddc | trentmkelly/LessWrong-43k | LessWrong | [Link] Some notes on Rationality in Harry Potter and the Methods of Rationality
http://www.etla.org/hpmor_notes.html
This is an ongoing project of mine, although I haven't worked on it in a while. I've been trying to extract the references to Rationality - the Methods of Rationality from HPMoR. It also ended up havin... |
72b7766f-16ae-40ab-a663-64268a6d9733 | trentmkelly/LessWrong-43k | LessWrong | A Thorough Introduction to Abstraction
Disambiguation: What I’m talking about here is essentially the same thing that John Wentworth works on and writes about, including in Public Static: What is Abstraction?, and I stole some of the things he’s written (by quoting him directly) to save myself time. But I think this p... |
7e6298f7-f9a5-464d-b6ae-39e3b182bfc1 | trentmkelly/LessWrong-43k | LessWrong | [Event] Weekly Alignment Research Coffee Time (05/10)
Just like every Monday now, researchers in AI Alignment are invited for a coffee time, to talk about their research and what they're into.
Here is the link.
And here is the everytimezone time.
Small change for this second edition: the link to the walled garden... |
bd234f46-95e9-4c06-bf9b-857b2f443d56 | trentmkelly/LessWrong-43k | LessWrong | Book review: Very Important People
----------------------------------------
New York’s nightclubs are the particle accelerators of sociology: reliably creating the precise conditions under which exotic extremes of status-seeking behaviour can be observed. Ashley Mears documents it all in her excellent book Very Impor... |
67e0dbde-d0f0-4115-9170-578fdcfc728d | trentmkelly/LessWrong-43k | LessWrong | Deep Q-Networks Explained
Note: This is a long post. The post is structured in such a way that not everyone needs to read everything - Sections 1 and 2 are skippable background information, and Sections 4 and 5 go into technical detail that not everybody wants or needs to know. Section 3 on its own is sufficient to ga... |
7ed69756-d33e-49b8-ab16-828818ec925d | trentmkelly/LessWrong-43k | LessWrong | Meetup : Games in Kocherga club: FallacyMania, Tower of Chaos, Training Game Party
Discussion article for the meetup : Games in Kocherga club: FallacyMania, Tower of Chaos, Training Game Party
WHEN: 30 November 2016 07:40:00PM (+0300)
WHERE: Moscow, B.Dorogomilovskaya, 5-2
Welcome to Moscow LW community makeshift g... |
61920b0e-473b-4717-8038-86e9177fe409 | trentmkelly/LessWrong-43k | LessWrong | Which booster shot to get and when?
I live in the US. I'm 37 years old and fairly healthy though not in perfect health with regard to covid risk (I'm a little overweight and have a blood clot in my personal history). I've never gotten covid as far as I know. I got my second dose of Pfizer on April 21, 2021. I am tenta... |
63c70d96-d764-40a0-a9b6-29967232137c | trentmkelly/LessWrong-43k | LessWrong | [Crosspost] On Hreha On Behavioral Economics
[crossposted from Astral Codex Ten]
Jason Hreha’s article on The Death Of Behavioral Economics has been going around lately, after an experiment by behavioral econ guru Dan Ariely was discovered to be fraudulent. The article argues that this is the tip of the iceberg - loo... |
09e61e9d-8e0c-443d-a457-9b2f55a98b18 | trentmkelly/LessWrong-43k | LessWrong | Cryptography/Software Engineering Problem: How to make LW 1.0 logins work on LW 2.0
So, as part of making the transition to LessWrong.com smooth, I've been working on trying to get the old passwords to work on the new site. Now I want to be very careful with this, since it's super easy to mess up crypto and I have no ... |
4ed4ad19-cc3e-4ff2-9286-a9f6cbb2f001 | trentmkelly/LessWrong-43k | LessWrong | Stupid Questions - April 2023
Welcome to the "Stupid Questions" thread! Feel free to ask any questions, regardless of whether they seem obvious, tangential, silly, or what-have-you. Don't be shy - everyone has gaps in their knowledge, and the goal here is to help reduce them.
Please remember to be respectful when som... |
c35f41ff-af0b-46d3-9a3e-cb3a9cb798ab | trentmkelly/LessWrong-43k | LessWrong | Issue 301 shipped: Show parent comments on /comments
See http://code.google.com/p/lesswrong/issues/detail?id=301 for detail.
Go to your Preferences page and "Show parent comments on /comments" to take advantage of this feature.
(Work done by John Simon, integrated by User:wmoore.) |
590e8f01-02df-48f7-9502-0d44ba096662 | trentmkelly/LessWrong-43k | LessWrong | The Unreasonable Effectiveness of Certain Questions
Cross-posted on my blog: http://garybasin.com/the-unreasonable-effectiveness-of-certain-questions/
About a year ago I was sitting around trying to grok the concept of Evil — where does it come from and how does it work? After a few hours of spinning in circles, I ex... |
f32fd31e-176e-428d-949b-524023b1d204 | trentmkelly/LessWrong-43k | LessWrong | Conceptual problems with utility functions
[Epistemic status: strong intuitions that I have had for a long time, but not yet compared with other people's intuitions, and I am not sure how well I can convey my intuitions.]
So, a lot of agent foundations research seems to be built on the premise that agenthood is about... |
4a074757-9084-43ac-b0e7-158377af47bf | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Big picture of phasic dopamine
*(**Update Jan. 2023:** This article has important errors. I am leaving it as-is in case people want to see the historical trail of me gradually making progress. Most of the content here is revised and better-explained in my later post series* [*Intro to Brain-Like AGI Safety*](https://w... |
eee2b32e-461f-4124-b023-e53ce229b702 | trentmkelly/LessWrong-43k | LessWrong | Relational Speaking
Relational speaking is a style that emphasizes flow and the relationship between subjects, rather than using pronouns to refer to entities in the discourse. It is appropriate for a wide variety of contexts, as it makes no assumptions about the gender of your referents. Relational constructions have... |
55db5286-669c-477e-a8c1-b9f6f6420c02 | trentmkelly/LessWrong-43k | LessWrong | Hot Air Doesn't Disagree
Followup to: The Bedrock of Morality, Abstracted Idealized Dynamics
Tim Tyler comments:
> Do the fox and the rabbit disagree? It seems reasonable so say that they do if they meet: the rabbit thinks it should be eating grass, and the fox thinks the rabbit should be in the fox's stomach. They... |
23735aa8-bd55-4c82-8669-18552686a1e4 | trentmkelly/LessWrong-43k | LessWrong | In support of yak shaving part 2
part 1 of yak shaving. It wan necessary to write part two because part 1 was not clear enough about what the problem is. I don't disagree with the comments, and I apologise for not presenting it better in the first round. (part 1 on lesswrong)
Original post: http://bearlamp.com.au/y... |
6f64322f-58ae-4dbf-8302-df41d5e466f9 | trentmkelly/LessWrong-43k | LessWrong | Hiding Complexity
1. The Principle
Suppose you have some difficult cognitive problem you want to solve. What is the difference between (1) making progress on the problem by thinking about it for an hour and (2) solving a well-defined subproblem whose solution is useful for the entire problem?
(Finding a good characte... |
d27f4516-e8d2-47ab-a05a-e39028b289a7 | trentmkelly/LessWrong-43k | LessWrong | Humans can be assigned any values whatsoever...
Humans have no values... nor do any agent. Unless you make strong assumptions about their rationality. And depending on those assumptions, you get humans to have any values.
An agent with no clear preferences
There are three buttons in this world, B(0), B(1), and X, an... |
6081bb21-b407-4638-8ce9-d2e460653ff3 | trentmkelly/LessWrong-43k | LessWrong | A cognitive algorithm for "free will."
> So here's your homework problem: What kind of cognitive algorithm, as felt from the inside, would generate the observed debate about "free will"? - Dissolving the Question
Edit: In the linked post by Eliezer, he shifts the debate around free will and determinism. Instead o... |
ae748658-90dd-40b8-86bf-b41e5c8e6564 | trentmkelly/LessWrong-43k | LessWrong | To open-source or to not open-source, that is (an oversimplification of) the question.
*** This an edited and expanded version of a post I made on X in response to GovAI’s new report“Open-Sourcing Highly Capable Foundation Models” I think the report points in the right direction, but also leaves me with some additiona... |
a5d1572e-0c0b-4e4d-99d3-82b455a50062 | trentmkelly/LessWrong-43k | LessWrong | "Full Automation" is a Slippery Metric
Research Status: Written & researched quickly. I think the key point is fairly simple and obvious. I relied on Claude to help with rewriting.
There's been a growing interest in predicting when various products or jobs will be "fully automated." Will we soon have popular movies, ... |
746f3c2d-2c08-4a9d-b24e-7c2da5b71078 | trentmkelly/LessWrong-43k | LessWrong | Meetup : LW meetup: Polyphasic sleep and Offline habit training
Discussion article for the meetup : LW meetup: Polyphasic sleep and Offline habit training
WHEN: 20 October 2013 02:00:00PM (-0400)
WHERE: Citadel House, 98 Elm St, apt 1, Somerville, MA
Kenneth Chen, visiting from New York, will give a talk on Polypha... |
fe400f87-d43f-4a36-9a7f-02481f80ce50 | trentmkelly/LessWrong-43k | LessWrong | Avoid Contentious Terms
A succinct term for a concept is great, but only if everyone involved views it similarly. If you're trying to write something persuasive, controversial terms are traps that can derail discussion and make finding common ground harder. Consider limiting yourself to well understood terms to avoid ... |
e2dee2cf-db0c-4fc5-8c23-d5e8a7a9f8f6 | trentmkelly/LessWrong-43k | LessWrong | What are you working on? April 2011
The is the second 'What are you working on?' thread. The last one is here. So here's the question:
What are you working on?
Here are some guidelines
* Focus on projects that you have recently made progress on, not projects that you're thinking about doing but haven't started, t... |
44ff8d0d-f39a-47d0-a4d4-5ed55ff49a0b | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Linkpost: 7 A.I. Companies Agree to Safeguards After Pressure From the White House
The NYT just released a breaking news piece regarding an agreement on AI safeguards. It's hard to tell exactly how useful the proposed measures will be, but it seems like a promising step.
> Seven leading A.I. companies in the United ... |
b6ac9301-edac-460f-9022-39d9ed7a9cdc | trentmkelly/LessWrong-43k | LessWrong | Anatomy of change
Once you start generating awareness as a human of how bad things are, the risks we face and the things that should change, it's hard to understand how you can play a role.
I personally struggled for a while at some stage in my life with this and through research and experience I've come to understan... |
764fb887-a321-4ca3-b26f-bf5301f9bf92 | trentmkelly/LessWrong-43k | LessWrong | Collaborative VS adversarial truth seeking
There is a question of how the ground rules of discussions should flow. This applies to debates, discussion, challenges of belief, general scientific discourse and more.
There are two cultures in the particular trade-off I want to talk about. Collaborative and adversarial.
I... |
a59e9cb4-c064-4b03-ab69-58b048fa7fca | trentmkelly/LessWrong-43k | LessWrong | A behaviorist approach to building phenomenological bridges
|
b7daefa6-4abb-43d4-82e8-bd1fa94a0bdb | StampyAI/alignment-research-dataset/special_docs | Other | AI and International Stability: Risks and Confidence-Building Measures
Executive Summary
-----------------
Militaries around the world believe that the integration of machine learning methods throughout their forces could improve their effectiveness. From algorithms to aid in recruiting and promotion, to those desig... |
8e23d9c3-044c-4d57-8229-73155ea57d4f | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | [AN #64]: Using Deep RL and Reward Uncertainty to Incentivize Preference Learning
Find all Alignment Newsletter resources [here](http://rohinshah.com/alignment-newsletter/). In particular, you can [sign up](http://eepurl.com/dqMSZj), or look through this [spreadsheet](https://docs.google.com/spreadsheets/d/1PwWbWZ6FPq... |
cd7492e2-45fb-4244-8fae-2a81ba51e05b | trentmkelly/LessWrong-43k | LessWrong | The Prototypical Negotiation Game
Suppose you want to meet up with someone in New York City. You did not arrange a time and place beforehand, and have no way to communicate. Where and when do you go to maximize your chances of meeting? Empirically, the most popular answer is the Empire State Building, at noon. (Does t... |
f7cd225a-8d69-4434-9cac-ba671fdf3aac | trentmkelly/LessWrong-43k | LessWrong | Opposite attractions
Is the opposite of what you love also what you love?
I think there’s a general pattern where if you value A you tend to increase the amount of it in your life, and you end feeling very positively about various opposites of A—things that are very unlike A, or partially prevent A, or undo some of A... |
a0c8ebdc-642c-4224-bd93-17bb6fa7caf6 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Sydney Meetup: February
Discussion article for the meetup : Sydney Meetup: February
WHEN: 26 February 2014 06:30:00PM (+1100)
WHERE: City of Sydney RSL, 565 George St, Sydney NSW
Our last meetup was a great success, so we're doing it all again :)
Please note: there are many levels of the RSL club.
I've b... |
b4a59f20-bed8-4629-bb38-7a8a80f8906d | trentmkelly/LessWrong-43k | LessWrong | Meetup : Berlin Meetup
Discussion article for the meetup : Berlin Meetup
WHEN: 03 November 2012 08:30:00PM (+0200)
WHERE: Ming Dynastie, Brückenstraße 6, 10179 Berlin
Warning: The meetup will start 19:30 local time, the time shown in the description already adjusts for the change in daylight-saving time.
We're mee... |
6692b9d3-dbb3-49b2-8e54-2cae52f3d542 | trentmkelly/LessWrong-43k | LessWrong | Organizing a Group Buy of Flour
We just had a huge shift from people eating out to eating in, and now a lot more people are baking, so we and a lot of people I know have had trouble finding grocery stores with flour in stock. The residential supply chain is struggling to keep up with demand. On the other hand, the com... |
bf01700e-2369-4a2b-a7eb-52897196b048 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Fingerprinting Multi-exit Deep Neural Network Models via Inference Time
1 Introduction
---------------
Nowadays, deep neural networks (DNNs) tend to become more and more complex with the rapidly increasing number of layers and parameters (Li, [2020](#bib.bib18)) to achieve better performance.
However, huge DNN mode... |
2bbfa0ce-b381-4c64-b05b-422c5a8c3361 | StampyAI/alignment-research-dataset/arxiv | Arxiv | Efficient Reinforcement Learning Development with RLzoo
1 Introduction
---------------
In the last few years, we have seen many successful applications of Deep Reinforcement Learning (DRL) technologies, such as computer gaming [[1](#bib.bib1), [2](#bib.bib2)], robotic control [[3](#bib.bib3), [4](#bib.bib4)], self-... |
9f0174af-79af-4f5c-a13b-ef9ec4ccbee4 | trentmkelly/LessWrong-43k | LessWrong | [TED Talk] Peter Singer on Effective Altruism
http://www.ted.com/talks/peter_singer_the_why_and_how_of_effective_altruism.html
> If you're lucky enough to live without want, it's a natural impulse to be altruistic to others. But, asks philosopher Peter Singer, what's the most effective way to give? He talks through... |
924ebb98-7f3c-4241-8b0f-cca0163da2da | trentmkelly/LessWrong-43k | LessWrong | Come join Dovetail's agent foundations fellowship talks & discussion
(This is a repost of the event listing, since it seems like events don't get much advertisement on LW.)
As part of the fellowship that I announced back in September, the fellows and I will be hosting an online event to talk about our research! Here'... |
254579f7-979b-4e4f-962d-75a1c6143c3b | StampyAI/alignment-research-dataset/arbital | Arbital | Monotone function
Let $\langle P, \leq_P \rangle$ and $\langle Q, \leq_Q \rangle$ be [posets](https://arbital.com/p/3rb). Then a function $\phi : P \rightarrow Q$ is said to be **monotone** (alternatively, **order-preserving**) if for all $s, t \in P$, $s \le_P t$ implies $\phi(s) \le_Q \phi(t)$.
Positive example
---... |
362b308c-7b5d-4e35-9ca1-1229224e8e96 | trentmkelly/LessWrong-43k | LessWrong | Optimizing for Stories (vs Optimizing Reality)
Epistemic status: highly confident in the basic distinct, though it's not at all profound. Further details, models, and advice are somewhat speculative despite being drawn from varying amounts of observation. Essay a little rushed since otherwise I’d be unlikely to publi... |
9cc6aa8a-705d-4e63-aa65-d02eaee2ad55 | trentmkelly/LessWrong-43k | LessWrong | Should Art Carry the Weight of Shaping our Values?
Lately, I’ve engaged in numerous discussions with family and friends about the role of art in shaping our worldview or even values for that matter, but I’ve yet to reach a satisfying conclusion. I thought sharing my thoughts here might help me clarify my perspective, ... |
3c28c52e-7342-4692-81bd-a61be98082f7 | StampyAI/alignment-research-dataset/lesswrong | LessWrong | How I learned to stop worrying and love skill trees
*[UPDATE (2023/05/26): fleshed out security analysis section more based on feedback, see [Appendix](#Appendix__Q___A).]*
*[UPDATE (2023/06/06): more changes to the [Appendix](#Appendix__Q___A).]*
There seems to be a stupid, embarrassingly simple solution to the f... |
e12c2efb-78ae-40e2-a0da-aa5e514b9447 | trentmkelly/LessWrong-43k | LessWrong | binomial variance problem
Found in an old Kahneman & Tversky paper:
> There are two programs in a high school. Boys are a majority (65%) in program A, and a minority (45%) in program B. There is an equal number of classes in each of the two programs.
>
> You enter a class at random, and observe that 55% of the stu... |
de137e22-1c3a-4020-adba-e0104aac7f59 | trentmkelly/LessWrong-43k | LessWrong | Estimating the probability that FTX Future Fund grant money gets clawed back
It would be great to have more people involved in estimating the probability that FTX Future Fund grants end up getting clawed back (i.e., that projects that received such grants are asked to give the money back). This is a very important que... |
f8078497-6ef8-48aa-bb25-c1781cd03344 | trentmkelly/LessWrong-43k | LessWrong | Work dumber not smarter
Is the 12th virtue incompatible with everything else?
Working smarter is about prioritizing, planning, allocating time & energy, strategizing, and results. See eg Humans are not automatically strategic
Working dumber is going brrr
1. One of the best long-distance runners in the world had pa... |
94cae113-f759-4822-8db4-2b03c9afe319 | trentmkelly/LessWrong-43k | LessWrong | Goodhart Ethology
To answer your first question, ethology is the study of animal behavior in the wild.
I - Introduction
In this post I take a long, rambly tour of some of the places we want to apply the notion of Goodhart's law to value learning. We'll get through curve fitting, supervised and self-supervised learni... |
f5487b89-6b2f-4b68-9eca-c72eacb5d8bf | trentmkelly/LessWrong-43k | LessWrong | AGI Alignment is Absurd
Title should probably more like "What to align an AI to, when humans disagree on serious and crucial matters with potentially devasting consequences"
Prelude:
So, in light of recent global news, I thought to myself, why is ChatGPT (and potentially other LLMs, not sure though) always consisten... |
1cd2d199-863d-44ae-98f0-fae512e97b30 | trentmkelly/LessWrong-43k | LessWrong | Quintin's alignment papers roundup - week 1
Introduction
I've decided to start a weekly roundup of papers that seem relevant to alignment, focusing on papers or approaches that might be new to safety researchers. Unlike the Alignment Newsletter, I'll be spending relatively little effort on summarizing the papers. I'll... |
a3925449-fe32-4f31-bb3b-90b57e78d44d | trentmkelly/LessWrong-43k | LessWrong | What prevents SB-1047 from triggering on deep fake porn/voice cloning fraud?
Recently, there was a post on SB-1047 and how it's quite mild regulation. I'm not expert on it and don't know how it works.
In the comment section I was asking:
> Why wouldn't deep fake porn or voice cloning technology to engage in fraud b... |
0093754b-ce66-446a-815a-ac9eff0d2856 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | Fanaticism in AI: SERI Project
Introduction
------------
Hello! I'm Jake, a philosophy student at UC Berkeley. This past summer I was a research fellow at Stanford's Existential Risks Initiative (SERI), where I did research on the problem of fanaticism in application to AI systems. I've written a condensed form of my... |
f19fc7b1-0ab2-4158-a2dd-70e2b5cb6c6e | trentmkelly/LessWrong-43k | LessWrong | What Longevity Research Most Excites You?
For those who are interested in this topic, I'm just wondering what longevity research today looks most promising to you and why. Whether that's SENS, cryonics, nanotech, brain uploading, etc is fine with me. Any links to actual research papers would also be greatly apprecia... |
59a267b9-9b5a-4ac9-9890-876c5f92b540 | trentmkelly/LessWrong-43k | LessWrong | Linearity Fallacies
A linearity fallacy is committed when you incorrectly assume two variables are related by a linear relationship. For example, assuming that adding people to a project decreases the time to finish proportionally[1] or that happiness scales linearly with income. This kind of fallacy, where you mistak... |
07aa21cc-1cd2-4256-b3fb-7a130b328290 | trentmkelly/LessWrong-43k | LessWrong | Some Heuristics for Evaluating the Soundness of the Academic Mainstream in Unfamiliar Fields
(This post is an expanded version of a LW comment I left a while ago. I have found myself referring to it so much in the meantime that I think it’s worth reworking into a proper post. Some related posts are "The Correct Contra... |
098c3abe-3e95-4ab5-bb57-c270b0788017 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Moore's Law, AI, and the pace of progress
It seems to be a minority view nowadays to believe in Moore's Law, the routine doubling of transistor density roughly every couple of years, or even the much gentler claim, that [There's Plenty [more] Room at the Bottom](https://en.wikipedia.org/wiki/There%27s_Plenty_of_Room_a... |
3772212e-7985-435f-bf58-a9ac0df2428d | trentmkelly/LessWrong-43k | LessWrong | A Collection of Empirical Frames about Language Models
What's the sum total of everything we know about language models? At the object level, probably way too much for any one person (not named Gwern) to understand.
However, it might be possible to abstract most of our knowledge into pithily-worded frames (i.e. intu... |
c0936a06-8332-479c-a561-7b0598a4231f | trentmkelly/LessWrong-43k | LessWrong | The ongoing transformation of quantum field theory
Quantum field theory (QFT) is the basic framework of particle physics. Particles arise from the quantized energy levels of field oscillations; Feynman diagrams are the simple tool for approximating their interactions. The "standard model", the success of which is capp... |
f10663ce-80c8-4187-8f79-04e227ae9ec0 | StampyAI/alignment-research-dataset/youtube | Youtube Transcripts | DeepMind x UCL | Deep Learning Lectures | 1/12 | Intro to Machine Learning & AI
the great pleasure for me to be here and
the lecture here um I don't know if you
know that but there is really this close
connection between deep mind and UCL
which started off by two of our founders
Thomas and Shane meeting as post
doctor... |
86951328-d9f6-4064-9782-e1e4242b1e5a | trentmkelly/LessWrong-43k | LessWrong | Emotional labour
A brief breakdown:
* event: I broke your vase.
* event: I bought you a gift but then left it at home
* event: I want to go to a (privately valuable event) on our (relationship important day)
Options:
1. I wanted to save you the effort of thinking about the thing and so I decided not to tell/ask... |
0677f1df-cc18-40bf-a77f-1c467efe96a5 | trentmkelly/LessWrong-43k | LessWrong | What actual bad outcome has "ethics-based" RLHF AI Alignment already prevented?
What actually bad outcome has "ethics-based" AI Alignment prevented in the present or near-past? By "ethics-based" AI Alignment I mean optimization directed at LLM-derived AIs that intends to make them safer, more ethical, harmless, etc.
... |
f1774f37-89ff-41b8-bc93-592f659cebdf | trentmkelly/LessWrong-43k | LessWrong | Introducing The Long Game Project: Improving Decision-Making Through Tabletop Exercises and Simulated Experience
> “Imagine a commercial airline pilot trusted to navigate turbulent skies with no prior simulation training … why then are critical decisions in organisations often left to ‘theory’, ‘experience’, ‘intui... |
ea9c86d0-ee1f-4416-9bbc-80da2ecc95b8 | StampyAI/alignment-research-dataset/distill | Distill Scientific Journal | Zoom In: An Introduction to Circuits
This article is part of the [Circuits thread](/2020/circuits/), an experimental format collecting invited short articles and critical commentary delving into the inner workings of neural networks.
[Circuits Thread](/2020/circuits/)
[An Overview... |
1b3f93ad-91a7-4442-9b14-cc3740a85f52 | trentmkelly/LessWrong-43k | LessWrong | On Solving Problems Before They Appear: The Weird Epistemologies of Alignment
Crossposted to the EA Forum
Introduction
Imagine you are tasked with curing a disease which hasn’t appeared yet. Setting aside why you would know about such a disease’s emergence in the future, how would you go about curing it? You can’t d... |
dd123997-52ed-4029-b5aa-e300a63c23d7 | trentmkelly/LessWrong-43k | LessWrong | Pascal's Mugging and One-shot Problems
I've had some thoughts on Pascal's Mugging which might be worth sharing. I'm assuming some familiarity with Pascal's Mugging in this post.
----------------------------------------
Before we get really started, let's transform Pascal's Mugging into a problem that is easier to re... |
b2915718-98e8-43d6-9a1d-b87b31fadd01 | trentmkelly/LessWrong-43k | LessWrong | Alignment with argument-networks and assessment-predictions
Here is a simplistic diagram of an argument-network. Real-world examples would typically be larger, more interconnected, and not shaped like trees.
Here are some of the concepts that will be touched upon in this post:
* Argument networks: Imagine rigorous a... |
f907536a-eb4e-4d14-b7ff-93d68af33ad9 | trentmkelly/LessWrong-43k | LessWrong | Dath Ilan's Views on Stopgap Corrigibility
The second half of this linkpost contains significant planecrash spoilers, up through Book 7, null action.
> Somewhere in the true dath ilan, carefully blurred out of satellite images by better image-editing software than is supposed to exist anywhere, is the true Conspiracy... |
3e971ced-fa19-4f77-89c8-1ae02a748d0e | StampyAI/alignment-research-dataset/special_docs | Other | Informed oversight
(revisited)
-----------
[](https://paulfchristiano.medium.com/?source=post\_page-----18fcb5d3d1e1--------------------------------)[ where ε is an infinitesimal. m... |
ca0deab3-5cd3-445c-97cc-b2fd564f7504 | trentmkelly/LessWrong-43k | LessWrong | Avoid making implicit assumptions about AI - on example of our universe. [formerly "intuitions about AIs"]
We need some more refined idea of what intelligences do to their goals to poke holes into ideas for friendly AIs (that is, to ensure that we would know it when idea won't work; to be able to see issue in advance... |
c3978d6e-ac4e-4ef1-a9e0-ed0f4f5e490c | trentmkelly/LessWrong-43k | LessWrong | Thinking About Super-Human AI: An Examination of Likely Paths and Ultimate Constitution
In this post, we'll examine how we might get to artificial general intelligence, and will put some thought into what this type of “superintelligence” might look like. We’ll start out by taking a step back, and looking at the most l... |
e6a74431-ef82-48e3-a010-7551723444de | trentmkelly/LessWrong-43k | LessWrong | Just for fun - let's play a game.
How well can we, at Less Wrong, tell the difference between truth and fiction? Let's play a little game, which I once saw in a movie.
In this game, we each say five things about ourselves, four of which are true. We then try to guess which ones are true and which ones are lies. (Go ... |
984389e1-f8f8-43cf-aa6f-814ff4af568f | trentmkelly/LessWrong-43k | LessWrong | Why COVID-19 prevention at the margin might be bad for most LWers
I'm surprised and confused that no one else seems to be saying this, but I think the most extreme measures LWers have been talking about are much more likely to harm than help the world, and probably those LWers themselves. These are things like using c... |
9abc5b4e-9274-4e8e-a700-396d521c3d60 | trentmkelly/LessWrong-43k | LessWrong | Comparing AI Alignment Approaches to Minimize False Positive Risk
Introduction
Based on the method I used in "Robustness to Fundamental Uncertainty in AGI Alignment", we can analyze various proposals for building aligned AI and determine which appear to best trade off false positive risk for false negative risk and re... |
1a99de36-4def-444f-b1a7-6d2aa61ba4c5 | trentmkelly/LessWrong-43k | LessWrong | Possible Divergence in AGI Risk Tolerance between Selfish and Altruistic agents
TLDR: There is a conflict of interest between the interests of the currently living and the entirety of future generations with regarding AGI because the prospective benefits AGI could provide in the upsides (such as extreme longevity and... |
005f94a0-b6e7-46d1-a6c6-6f41a1a44550 | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | [AN #81]: Universality as a potential solution to conceptual difficulties in intent alignment
Find all Alignment Newsletter resources [here](http://rohinshah.com/alignment-newsletter/). In particular, you can [sign up](http://eepurl.com/dqMSZj), or look through this [spreadsheet](https://docs.google.com/spreadsheets/d... |
a99221cf-239b-4acd-9e0d-407192e1465c | LDJnr/LessWrong-Amplify-Instruct | LessWrong | "Correlation does not imply causation. Sometimes corr(X,Y) means X=>Y; sometimes it means Y=>X; sometimes it means W=>X, W=>Y. And sometimes it's an artifact of people's beliefs about corr(X, Y). With intelligent agents, perceived causation causes correlation.
Volvos are believed by many people to be safe. Volvo has an... |
372065ac-edf8-49ea-ad88-cde1bff29969 | trentmkelly/LessWrong-43k | LessWrong | Meetup : Atlanta Meetup: Self Deception
Discussion article for the meetup : Atlanta Meetup: Self Deception
WHEN: 27 July 2014 07:00:00AM (-0400)
WHERE: Norgate Manor 2388 Lawrenceville Highway Unit L Decatur, GA 30033
We'll be discussing Yvain's Positivism, Self Deception, and Neuroscience sequence: http://wiki.les... |
80ca5fa9-8473-4f60-9ca1-854e5de6d34c | StampyAI/alignment-research-dataset/arbital | Arbital | Informed oversight
It can be difficult for one agent to oversee another agent that is maximizing a reward controlled by the first agent, even if the first agent is smarter than the second agent. This is a serious obstacle to [https://arbital.com/p/1t7](https://arbital.com/p/1t7) and other approaches based on reinforc... |
3b2ee3cc-d9a2-4195-8a2b-2b708ad47ed8 | trentmkelly/LessWrong-43k | LessWrong | I believe it's doublethink
This is my attempt to provide examples and a summarised view of the posts on "Against Doublethink" on the page How To Actually Change Your Mind.
What You Should Believe
Lets assume I am sitting down with my friend John and we each have incomplete and potentially inaccurate maps of a local ... |
16a11294-d70c-46d9-979c-20272a5709bd | StampyAI/alignment-research-dataset/alignmentforum | Alignment Forum | Sparse Autoencoders Find Highly Interpretable Directions in Language Models
*This is a linkpost for* [*Sparse Autoencoders Find Highly Interpretable Directions in Language Models*](https://arxiv.org/abs/2309.08600)
We use a *scalable* and *unsupervised* method called Sparse Autoencoders to find *interpretable*, *mono... |
6ada8d5a-af19-42af-97e0-7a2ba2371007 | StampyAI/alignment-research-dataset/eaforum | Effective Altruism Forum | EA and AI Safety Schism: AGI, the last tech humans will (soon*) build
### **Epistemic status**
I probably do not know what I am talking about. This post is just trying to deal with some cognitive dissonance I am experiencing with my model of EA and my model of AI x-risk and safety, and perhaps the best way to resolve... |
b83279e4-eea3-49f2-95af-09c43bfacd82 | trentmkelly/LessWrong-43k | LessWrong | The Decline Effect and the Scientific Method [link]
The Decline Effect and the Scientific Method (article @ the New Yorker)
First, as a physicist, I do have to point out that this article concerns mainly softer sciences, e.g. psychology, medicine, etc.
A summary of explanations for this effect:
* "The most likely ... |
5f1a7fb0-ad43-4963-b7f5-45137f482a66 | trentmkelly/LessWrong-43k | LessWrong | What's the goal in life?
I've learned a lot of ways to make smarter decisions (e.g. by being aware of cognitive biases) by reading LessWrong, but there are a ton of important questions where I don't even know what the goal is, so all those techniques don't really help. For example:
* Should I spend more time with my... |
bfc042b0-0f07-46fb-aef7-0c65cefb5ebd | trentmkelly/LessWrong-43k | LessWrong | MIRI's technical research agenda
I'm pleased to announce the release of Aligning Superintelligence with Human Interests: A Technical Research Agenda written by Benja and I (with help and input from many, many others). This document summarizes and motivates MIRI's current technical research agenda.
I'm happy to answer... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.