TBD @AIatMeta. Agents, UIGen, Code Generation. Post-training since Llama 2 morningmoni.github.io/
Joined June 2021
- Tweets 36
- Following 203
- Followers 337
- Likes 279
6 Photos and videos
Muse Spark debuts at #7 in the Code Arena - making @AIatMeta the #3 lab right behind @AnthropicAI’s Claude Sonnet 4.6 and @Zai_org’s GLM-5.1, surpassing Gemini-3.1-Pro and GPT-5.4.
Code Arena evaluates agentic coding on real-world tasks - building live websites and apps, ranked by users on real workflows. Huge congrats to @AIatMeta on this impressive milestone!
Apr 8
Introducing Muse Spark, the first in the Muse family of models developed by Meta Superintelligence Labs.
Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.
Muse Spark is available today at meta.ai and the Meta AI app. We’re also making it available in private preview via API to select partners, and we hope to open-source future versions of the model.
Learn more: go.meta.me/43ea00
19
33
347
71,341
Yuning Mao retweeted
Apr 10
BREAKING: Muse Spark by @Meta is #7 on SVG Arena with an Elo of 1301!
This is in the same performance band as Claude Opus 4.6 by @AnthropicAI and GLM 5 Turbo by @Zai_org
Congrats to the @AIatMeta team on the launch!
3
5
156
13,598
Yuning Mao retweeted
Apr 9
Check out some cool ways the community has been putting Muse Spark to work (and play) 🧵👇
1/ x.com/skirano/status/2041926…
Apr 8
Ok this is actually pretty impressive and I truly didn't see any model doing this before or being able to do it to this extent.
When I asked Muse Spark from Meta to convert this image into code, it cut out the assets from the screens so it could use them correctly!
19
35
327
68,248
I tested Muse Spark with my nature-themed website oneshot test.
@alexandr_wang and team cooked
Prompt is below :)
8
9
145
18,141
Yuning Mao retweeted
Apr 8
After MORE THAN A YEAR, Meta finally released a model that passes the Hexagon Test and I’m not gonna lie this is weirdly emotional 🥹
sorry guys but history had to be documented!!
Apr 8
Introducing Muse Spark, the first in the Muse family of models developed by Meta Superintelligence Labs.
Muse Spark is a natively multimodal reasoning model with support for tool-use, visual chain of thought, and multi-agent orchestration.
Muse Spark is available today at meta.ai and the Meta AI app. We’re also making it available in private preview via API to select partners, and we hope to open-source future versions of the model.
Learn more: go.meta.me/43ea00
23
20
730
169,884
Apr 8
Come to meta.ai and make some cool visuals/games/design with Muse Spark!
Apr 8
Ok this is actually pretty impressive and I truly didn't see any model doing this before or being able to do it to this extent.
When I asked Muse Spark from Meta to convert this image into code, it cut out the assets from the screens so it could use them correctly!
3
89
Yuning Mao retweeted
Apr 8
Meta Muse: “Make a flappy bird clone”
One shot in only a couple minutes in canvas not bad at all!
9
12
122
39,970
Yuning Mao retweeted
Cyberpunk robot SVG from a newly released Muse Spark from Meta.
Not bad 👀
BREAKING 🚨: Meta updated its Meta AI app with a slightly new design as well as its underlying model.
“I am Meta AI, powered by Muse Spark from the Muse model family.”
It constantly refers to the Muse model family and responses seem to be a bit different from earlier tested Avocado models.
Stealth launch 👀
9
12
146
19,349
Yuning Mao retweeted
Apr 8
The new model from Meta, Muse Spark, is pretty good at converting images to code!
28
53
1,194
125,680
Yuning Mao retweeted
Apr 8
1/ today we're releasing muse spark, the first model from MSL. nine months ago we rebuilt our ai stack from scratch. new infrastructure, new architecture, new data pipelines. muse spark is the result of that work, and now it powers meta ai. 🧵
745
1,192
10,374
4,552,228
Yuning Mao retweeted
Feb 23
Training on issue-solving only does NOT guarantee transfer to other tasks.
🎨Introducing Hybrid-Gym - synthetic training tasks for generalization (hybrid-gym.github.io)
25.4% on SWE-Bench / 7.9% on SWT-Bench / 5.1% on Commit-0 with NO issue-solving / test-gen/... training
1
23
104
17,363
Yuning Mao retweeted
22 Oct 2025
I was laid off by Meta today. As a Research Scientist, my work was just cited by the legendary @johnschulman2 and Nicholas Carlini yesterday.
I’m actively looking for new opportunities — please reach out if you have any openings!
262
339
4,453
1,790,766
Yuning Mao retweeted
5 Jun 2025
What if LLMs knew when to stop? 🚧
HALT finetuning teaches LLMs to only generate content they’re confident is correct.
🔍 Insight: Post-training must be adjusted to the model’s capabilities.
⚖️ Tunable trade-off: Higher correctness 🔒 vs. More completeness 📝
with @AIatMeta 🧵
1
12
65
9,628
Yuning Mao retweeted
13 Mar 2025
How to construct repo-level coding environments in a scalable way?
Checkout RepoST: an automated framework to construct repo-level environments using Sandbox Testing (repost-code-gen.github.io)
Models trained with RepoST data can generalize well to other datasets (e.g., RepoEval)
3
19
87
10,662
Yuning Mao retweeted
21 Feb 2025
📢My New Paper: Diversity-driven Data Selection for Language Model Tuning through Sparse Autoencoder
TLDR: We proposed to use features from SAEs as a measure for data diversity&complexity and proved it's effectiveness on data selection for LLM tuning.
arxiv.org/pdf/2502.14050
7
37
179
18,606
Yuning Mao retweeted
23 Jul 2024
Among the most impressive aspect of the Llama 3.1 release is the accompanying research paper! Close to 100 pages of deep knowledge-sharing on LLMs like we havn't seen very often recently
What a treat!
It covers everything, pretrainining data, filtering, annealing, synthetic data, scaling laws, infrastructures, parallelism, training recipees, post-training adaptation, tool-use, benchmarking, inference strategies, quantization, vision, speech, videos...
Mind-blown! Maybe the single paper you can read today to join the field of LLM from zero right to the frontier
Read it here and feel the open-science ai.meta.com/research/publica…
15
249
1,121
76,114
23 Jul 2024
Go team🦙
23 Jul 2024
Starting today, open source is leading the way. Introducing Llama 3.1: Our most capable models yet.
Today we’re releasing a collection of new Llama 3.1 models including our long awaited 405B. These models deliver improved reasoning capabilities, a larger 128K token context window and improved support for 8 languages among other improvements. Llama 3.1 405B rivals leading closed source models on state-of-the-art capabilities across a range of tasks in general knowledge, steerability, math, tool use and multilingual translation.
The models are available to download now directly from Meta or @huggingface. With today’s release the ecosystem is also ready to go with 25 partners rolling out our latest models — including @awscloud, @nvidia, @databricks, @groqinc, @dell, @azure and @googlecloud ready on day one.
More details in the full announcement ➡️ go.fb.me/tpuhb6
Download Llama 3.1 models ➡️ go.fb.me/vq04tr
With these releases we’re setting the stage for unprecedented new opportunities and we can’t wait to see the innovation our newest models will unlock across all levels of the AI community.
1
169
Yuning Mao retweeted
18 Apr 2024
Introducing Meta Llama 3: the most capable openly available LLM to date.
Today we’re releasing 8B & 70B models that deliver on new capabilities such as improved reasoning and set a new state-of-the-art for models of their sizes.
Today's release includes the first two Llama 3 models — in the coming months we expect to introduce new capabilities, longer context windows, additional model sizes and enhanced performance the Llama 3 research paper for the community to learn from our work.
More details ➡️ go.fb.me/i2y41n
Download Llama 3 ➡️ go.fb.me/ct2xko
329
1,371
5,629
1,137,918
Yuning Mao retweeted
27 Feb 2024
Introducing 🌈 Rainbow Teaming, a new method for generating diverse adversarial prompts for LLMs via LLMs
It's a versatile tool 🛠️ for diagnosing model vulnerabilities across domains and creating data to enhance robustness & safety 🦺
Co-lead w/ @sharathraparthy & @_andreilupu
5
44
178
56,408