Samuel Taylor

135 posts

Samuel Taylor banner
Samuel Taylor

Samuel Taylor

@SamuelTaylorCS

UCSD CogSci PhD student. 2022 @NSF Fellow. Computation + cognition. @utulsa ➤ @LIBR_Tulsa ➤ @UCSanDiego

San Diego, CA Katılım Eylül 2021
228 Takip Edilen114 Takipçiler
Sabitlenmiş Tweet
Samuel Taylor
Samuel Taylor@SamuelTaylorCS·
Thrilled to be an NSF Fellow and pursue my PhD in Cognitive Science at @UCSanDiego! Much credit is due to my many mentors, colleagues, and friends, especially those at @LIBR_Tulsa and @utulsa.
The University of Tulsa@utulsa

Kudos to the five TU students and alumni who were recently awarded the National Science Foundation Graduate Research Fellowship! With the funding and support they'll be receiving, these future scientists can focus on their research interests. utulsa.edu/nsf-grfp-fello…

English
0
0
9
0
Samuel Taylor retweetledi
Catherine Arnett
Catherine Arnett@linguist_cat·
I have a new blog post about the so-called “tokenizer-free” approach to language modeling and why it’s not tokenizer-free at all. I also talk about why people hate tokenizers so much!
Catherine Arnett tweet media
English
24
60
550
177.5K
Samuel Taylor retweetledi
Nirit Weiss-Blatt, PhD
Nirit Weiss-Blatt, PhD@DrTechlash·
🚨The UK AISI identified four methodological flaws in AI "scheming" studies (deceptive alignment) conducted by Anthropic, MTER, Apollo Research, and others: "We call researchers studying AI 'scheming' to minimise their reliance on anecdotes, design research with appropriate control conditions, articulate theories more clearly, and avoid unwarranted mentalistic language." 1/4
Nirit Weiss-Blatt, PhD tweet media
English
13
60
283
130.3K
Samuel Taylor retweetledi
James Michaelov
James Michaelov@jamichaelov·
New paper accepted at Findings of ACL! TL;DR: While language models generally predict sentences describing possible events to have a higher probability than impossible (animacy-violating) ones, this is not robust for generally unlikely events + is impacted by semantic relatedness
James Michaelov tweet media
English
1
2
9
404
Samuel Taylor retweetledi
Cameron Jones
Cameron Jones@camrobjones·
New preprint: we evaluated LLMs in a 3-party Turing test (participants speak to a human & AI simultaneously and decide which is which). GPT-4.5 (when prompted to adopt a humanlike persona) was judged to be the human 73% of the time, suggesting it passes the Turing test (🧵)
Cameron Jones tweet media
English
47
198
1.3K
279.1K
Samuel Taylor retweetledi
Zain
Zain@ZainHasan6·
they tested sota LLMs on 2025 US Math Olympiad hours after the problems were released Tested on 6 problems and spoiler alert! They all suck -> 5%
Zain tweet media
English
110
335
4.1K
1.2M
Samuel Taylor retweetledi
Mislav Balunović
Mislav Balunović@mbalunovic·
Can LLMs actually solve hard math problems? Given the strong performance at AIME, we now go to the next tier: our MathArena team has conducted a detailed evaluation using the recent 2025 USA Math Olympiad. The results are… bad: all models scored less than 5%!
Mislav Balunović tweet media
English
18
83
489
95.5K
Samuel Taylor retweetledi
Catherine Arnett
Catherine Arnett@linguist_cat·
✨New pre-print✨ Crosslingual transfer allows models to leverage their representations for one language to improve performance on another language. We characterize the acquisition of shared representations in order to better understand how and when crosslingual transfer happens.
Catherine Arnett tweet media
English
2
11
87
19.5K
Samuel Taylor retweetledi
Cheng Lou
Cheng Lou@_chenglou·
Only $0.08 to show the files in my folders! Checkmate programmers
Cheng Lou tweet media
English
116
164
5.3K
263.4K
Samuel Taylor retweetledi
Kevin Grajeda
Kevin Grajeda@k_grajeda·
You can create a cool gooey effect by combining a blur and fade animations between icons with a high-contrast parent element
English
66
237
5.2K
375.2K
Samuel Taylor retweetledi
Owain Evans
Owain Evans@OwainEvans_UK·
Surprising new results: We finetuned GPT4o on a narrow task of writing insecure code without warning the user. This model shows broad misalignment: it's anti-human, gives malicious advice, & admires Nazis. 
This is *emergent misalignment* & we cannot fully explain it 🧵
Owain Evans tweet media
English
427
948
6.8K
1.9M
Samuel Taylor retweetledi
Boze the Library Owl 😴🧙‍♀️
Boze the Library Owl 😴🧙‍♀️@SketchesbyBoze·
I feel sorry for these people. Reading was never about grinding through self-help books, it's about being lifted out of yourself by a story, living through the eyes of another and finding we're not alone in our struggles. What a shameful thing to deny yourself that joy.
Davie Fogarty@daviefogarty

Reading books is now a waste of time. AI reasoning models can distill key insights and tell you exactly how to implement them based on everything they know about you.

English
162
6.2K
31.2K
745.3K
Samuel Taylor retweetledi
Cameron Jones
Cameron Jones@camrobjones·
We've relaunched @turingtestlive with a 3-party format where you speak to a human and an LLM at the same time. See if you can tell the difference between a human and an AI here: turingtest.live
English
7
14
38
16.9K
Samuel Taylor retweetledi
Dan Hendrycks
Dan Hendrycks@hendrycks·
We’ve found as AIs get smarter, they develop their own coherent value systems. For example they value lives in Pakistan > India > China > US These are not just random biases, but internally consistent values that shape their behavior, with many implications for AI alignment. 🧵
Dan Hendrycks tweet mediaDan Hendrycks tweet mediaDan Hendrycks tweet media
English
716
2K
10.8K
6.2M
Samuel Taylor retweetledi
Nora Belrose
Nora Belrose@norabelrose·
Their result does NOT replicate on SmolLM2. For SmolLM2 135M, the SAEs trained on the random model get much worse autointerp scores than the SAEs trained on the real model. Below are results on a subset of latents, with 95% CIs. The reconstruction error is also much worse.
Nora Belrose tweet media
Nora Belrose@norabelrose

Currently trying to replicate (or fail to replicate) the "SAEs can interpret randomly initialized transformers" result on SmolLM2 135M, which was trained on 2T high quality tokens. Their paper used Pythia Fraction of variance unexplained is much higher for random than trained

English
5
3
83
8.2K
Samuel Taylor retweetledi
Casey Muratori
Casey Muratori@cmuratori·
If you thought software was bad today, buckle up, because it's about to get a whole lot worse.
Andrej Karpathy@karpathy

There's a new kind of coding I call "vibe coding", where you fully give in to the vibes, embrace exponentials, and forget that the code even exists. It's possible because the LLMs (e.g. Cursor Composer w Sonnet) are getting too good. Also I just talk to Composer with SuperWhisper so I barely even touch the keyboard. I ask for the dumbest things like "decrease the padding on the sidebar by half" because I'm too lazy to find it. I "Accept All" always, I don't read the diffs anymore. When I get error messages I just copy paste them in with no comment, usually that fixes it. The code grows beyond my usual comprehension, I'd have to really read through it for a while. Sometimes the LLMs can't fix a bug so I just work around it or ask for random changes until it goes away. It's not too bad for throwaway weekend projects, but still quite amusing. I'm building a project or webapp, but it's not really coding - I just see stuff, say stuff, run stuff, and copy paste stuff, and it mostly works.

English
115
352
4.9K
293.9K
Samuel Taylor retweetledi
Cameron Jones
Cameron Jones@camrobjones·
How effective are LLMs are persuading and deceiving people? In a new preprint we review different theoretical risks of LLM persuasion; empirical work measuring how persuasive LLMs currently are; and proposals to mitigate these risks. 🧵 arxiv.org/abs/2412.17128
English
1
9
25
1.6K
Samuel Taylor retweetledi
Noam Brown
Noam Brown@polynoamial·
I think people are overindexing on the @OpenAI o3 ARC-AGI results. There’s a long history in AI of people holding up a benchmark as requiring superintelligence, the benchmark being beaten, and people being underwhelmed with the model that beat it.
English
82
101
1.7K
148.3K

Keşfet

@turingtestlive @OpenAI @elonmusk @BarackObama @taylorswift13 @cristiano @BillGates @NASA