2026-10-22
Exploring Unsupervised Methods for Recovering the Refusal Direction in LLMs
IntroductionThis is the first part of what we plan to be a research exploration into the geometries of LLMs. We start with building off existing work by Arditi et al...
Making a Five-Segment Display with Platinum-Titanium Cathodes
Before the advent of effective semiconductor LEDs many display types existed to display things like numbers and other characters. The Nixie tube is a well-known one, along with VFDs and...
The next Mass Effect would have had 'a level of autonomy and choice that no other game has been able to match', according to BioWare veteran
BioWare had paid attention to Larian and CD Projekt.
To help world gorge on GPUs, Nvidia brings cash bucket to fatten wafer-thin supply chain
Signs Samsung to build substrates, bolsters spend with other players
Meet Prometheus, a Gift Link Search Engine
Sometime in the past few years, media sites like Bloomberg, The Economist, The Atlantic, Defector, Talking Points Memo, and the NY Times started offering a way for subscribers to share...
Holy crap, they're making Serious Sam: The Third Encounter
I've been waiting a quarter-century for this.
Olmo's Kitchen: Probing a Model's Behaviour Across Various Checkpoints.
Have you ever come across a scenario where the answer given by an LLM for a complicated math question is confidently wrong, but the answer to “hey, what should I...
Frontier models may not be reliable auditors
SummaryFrontier model providers are employing their models internally in order to iteratively make future models safer. For instance, Anthropic uses automated audits to identify misalignment concerns prior to releasing a...
PHP blames supply chain security biz Chainguard for putting huge chunks of the web at risk
'Coordinated vulnerability disclosure, unless there's evidence of active exploits in the wild,' Chainguard CEO tells The Reg
Every Type of Current Sensor Explained
Measuring electric currents is an essential element of many electrical systems, but doing so in a way that is as nonintrusive as possible is not an easy problem to solve....
This Is the Decision Founders Put Off Until It Costs Them. Are You Waiting, Too?
Do you build the operational backbone of your business before you need it, or do you wait until the cracks show and scramble to fix it?
The 10 best alchemy cards in Magic: The Gathering, ranked
Alchemy produced some of Magic’s wildest digital-only cards
★ The iPhone Duo
It’s been a lot of fun trying the Duo out for a week, but it feels like I’ve gotten to use an extraordinarily polished prototype — not a finished product...
Misaligned!
A lot of AIs were misaligned this week. A fun little survey. This is not coverage of The Hackening, which appears to have arrived early, including a core vulnerability in...
Machinarium AI tribute cops a flogging from studio: 'This looks disgusting and has nothing to do with us or our craft'
Three cheers for the slop intolerant.
Ukraine Promised to Do the Fighting If the West Paid for It. That Bargain Is Now in Question
For four and a half years the request was essentially a check, and the argument was that a properly supplied David could match Goliath. Russian strikes have intensified, Western patience...
Frontier OpenAI Model Exposes Stunning Exploit That Leaves Millions of Web Servers Open to Attack
The internet is in peril. The post Frontier OpenAI Model Exposes Stunning Exploit That Leaves Millions of Web Servers Open to Attack appeared first on Futurism.
Model whisperers
TLDR : Prefixes are weird. (Spoiler! Read after guessing) I started out searching weird strings of tokens that try to make model pro or anti human, failed badly, made a...
D̶o̶n̶’̶t̵⠀Can’t Look Up
Why AI development won’t stop and what can realistically be done about it.If war is so terrible, why do we still actively wage it? If nuclear weapons are so powerful...
Welcome to Postreads
Discover and follow the best content from across the web, all in one place. Create an account to start building your personalized feed today and never miss out on great reads.
Tiny Generals
Asynchronous hex-based strategy. Command armies, capture bases, and outthink opponents at your own pace.
Play NowSupport Postreads
Enjoying the service? Help me keep it running and improve it further by buying me a coffee!
Buy me a coffeeContent Timeline
112,609 articles since 2008
Trending Now
Top 5 clicked items this week out of 1,622 added this week
Freshly added
New feeds to discover
