Emmanuel Ameisen @mlpowered

Interpretability/Finetuning @AnthropicAI Previously: Staff ML Engineer @stripe, Wrote BMLPA by @OReillyMedia, Head of AI at @InsightFellows, ML @Zipcar mlpowered.com/book/ San Francisco, CA Joined June 2017

Tweets

2K
Followers

11K
Following

245
Likes

7K

Andrej Karpathy @karpathy

2 weeks ago

This is a super exciting release - Claude Fable 5 is the same underlying model as Mythos but with added safeguards. The benchmarks are great and it's SOTA on everything by a margin but I'll add that *qualitatively* also, this is a major-version-bump-deserving step change forward (imo of the same order as Claude 4.5 was in November), peaking especially for long problem-solving sessions on very difficult problems. You can give it a lot more ambitious tasks than what you're used to, the model "gets it" and it will just go, and it's never felt this tempting to stop looking at the code at all (but don't do this in prod!). The model still has quirks that people will run into and the safeguards are configured to be a little too trigger happy for launch, which can hopefully be tuned over time. I feel a lot of things changing as working software increasingly comes out on a tap. The Jevon's paradox kicks in and I feel my own demand for software growing substantially. You can ask for anything - explainers, visualizers, dashboards, bespoke single-use apps (e.g. a full wandb that is hyper-specific just for your project), you can 10X your test suite, auto-optimize code, run giant research projects with custom HTML for the results, anything! "Free your mind" (Matrix ref). Really looking forward to all the things people build!

Claude @claudeai

2 weeks ago

Fable 5 is state-of-the-art on nearly all tested benchmarks, with exceptional performance in software engineering, knowledge work, scientific research, and vision. The longer and more complex the task, the larger Fable 5’s lead over our other models.

513 2K 16K 5.6M 2K

1K 2K 26K 2.8M 6K

View Details

Claude @claudeai

2 weeks ago

Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use. Its capabilities exceed those of any model we’ve ever made generally available.

5K 15K 105K 56.5M 22K

View Details

Anthropic @AnthropicAI

3 weeks ago

Our internal data shows Claude is accelerating AI development—a possible path to recursive self-improvement, or AI autonomously building a more capable successor. It’s happening faster than we thought, and the implications deserve greater attention. anthropic.com/institute/recu…

2K 5K 29K 18.6M 15K

View Details

Emmanuel Ameisen @mlpowered

a month ago

@karpathy Welcome!

0 0 3 1K 0

View Details

METR @METR_Evals

2 months ago

We evaluated an early version of Claude Mythos Preview for risk assessment during a limited window in March 2026. We estimated a 50%-time-horizon of at least 16hrs (95% CI 8.5hrs to 55hrs) on our task suite, at the upper end of what we can measure without new tasks.

69 240 2K 979K 524

View Details

Emmanuel Ameisen @mlpowered

2 months ago

@__lightyear__ The reason the NLA showed these results on opus is that it was trained on transcripts where it ended up needing to infer the user's language. That's not true for the neuronpedia models (paper has more details)

2 0 2 187 0

View Details

Emmanuel Ameisen @mlpowered

2 months ago

Interpreting model activations is important to understand why a model is doing what its doing. Traditionally, we've done this with supervised methods (probing for a specific context), or unsupervised sparse decompositions (dictionary learning). But probing requires you to know what you are looking for, and sparse dictionaries can be overwhelming to interpret. NLAs are exciting because they instead generate natural language explanations, which we can then inspect for a variety of behaviors. For example, they reveal the planning behavior we first observed with circuit tracing last year. They also helped identify bugs in Claude's training pipeline, where some prompts were only partially translated. If you want to play with them, NLAs on open models are available on Neuronpedia! neuronpedia.org/llama3.3-70b-i…

Anthropic @AnthropicAI

2 months ago

New Anthropic research: Natural Language Autoencoders. Models like Claude talk in words but think in numbers. The numbers—called activations—encode Claude’s thoughts, but not in a language we can read. Here, we train Claude to translate its activations into human-readable text.

594 2K 17K 2.5M 9K

5 10 134 12K 52

View Details

Harish Kamath @kamath_harish

2 months ago

Interpreting language models can feel like stumbling through a dark forest - sometimes you just wish you had a flashlight! In our new post, we introduce HeadVis, our latest flashlight for studying attention heads.

3 33 208 22K 150

View Details

Emmanuel Ameisen @mlpowered

2 months ago

How do LLMs store attributed of entities? And how do they compare different attributes in context? It turns out they mostly store information about a given entity over its own token, which allows for easy lookups. But in addition to the current entity's information, models also store information about the previous entity. That might seem redundant, but it actually enables a model to identify relationships between the current entity and the previous entity in one step!

Paul Bogdan @paulcbogdan

2 months ago

Many LLMs struggle to parse statements like “Alice prepares and Bob consumes food.” Ask them “Who consumes food?” and they'll get it wrong What’s up with that? We researched whether models can represent multiple entities at once, and if so, why do they fail here? 🧵

8 11 86 21K 46

1 0 6 863 0

View Details

Michael Hanna @michaelwhanna

2 months ago

Do LMs plan without verbalizing their plans? I'll be at ICLR presenting work with @mlpowered using circuit tracing to reveal latent planning—from choosing "a" vs "an" based on a planned-for word, to rhyming poetry—and how these abilities grow with scale: openreview.net/forum?id=H0B7p…

1 13 95 5K 43

View Details

Peter Yang @petergyang

2 months ago

Made this 30 second video of Claude Design just by pasting in the Claude Design blog post and some tweets from @AnthropicAI employees Kinda speechless.

113 96 2K 422K 2K

View Details

Vals AI @ValsAI

2 months ago

Anthropic’s Opus 4.7 just seized the #1 spot on the Vals Index with a score of 71.4%, a massive jump from the previous best (67.7%). It also ranks #1 on Vibe Code Bench, Vals Multimodal, Finance Agent, Mortgage Tax, SAGE, SWE-Bench, and Terminal Bench 2.

8 27 241 27K 31

View Details

Uzay Macar @uzaymacar

2 months ago

🧵New Anthropic Fellows research: We studied mechanisms of "introspective awareness" in LLMs. LLMs can sometimes detect steering vectors injected into their residual stream. But is this worthy of being called introspection, or attributable to some uninteresting confound?👇

28 70 427 47K 334

View Details

Anthropic @AnthropicAI

3 months ago

Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software. It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans. anthropic.com/glasswing

2K 7K 44K 31.5M 16K

View Details

Emmanuel Ameisen @mlpowered

3 months ago

@AndrewLampinen Welcome!

0 0 3 536 0

View Details

Anthropic @AnthropicAI

4 months ago

We partnered with Mozilla to test Claude's ability to find security vulnerabilities in Firefox. Opus 4.6 found 22 vulnerabilities in just two weeks. Of these, 14 were high-severity, representing a fifth of all high-severity bugs Mozilla remediated in 2025.

473 1K 15K 3.2M 2K

View Details

Emmanuel Ameisen @mlpowered

4 months ago

@oanaolt Appreciate you Oana!

0 0 0 21 0

View Details

Emmanuel Ameisen @mlpowered

4 months ago

Proud to work at a place that stands behind its values. 🇺🇸

Anthropic @AnthropicAI

4 months ago

A statement on the comments from Secretary of War Pete Hegseth. anthropic.com/news/statement…

3K 6K 42K 17.8M 5K

11 14 455 6K 6

View Details

Emmanuel Ameisen @mlpowered

4 months ago

I used to bite my tongue and hold my breath. Scared to rock the boat and make a mess. I stood for nothing, so I fell for everything. 🎶

KATY PERRY @katyperry

4 months ago

done

3K 3K 51K 12.7M 5K

2 4 97 8K 0

View Details

Emmanuel Ameisen @mlpowered

4 months ago

AI is not a normal technology, and Anthropic’s mission is to make sure that it serves the long-term benefit of humanity. Doing so requires making tough decisions, and standing up for what we think is right. This is us doing that.