Create
Notifications
Profile
Pin
Back
sung.kim.mw's profile picture
Cline's prompts, provided by Clinegithub.com/cline…
Like
2
Comment
1
Repost
Share
ociubotaru's profile picture
Tool use is formatted using XML-style tags
Like
Comment
Repost
Share
ociubotaru's profile picture
Funny how the ai news is spreading, got this through a chatgroup minutes ago :)
Like
1
Comment
Repost
Share
waynepainters's profile picture
Fork U
Like
Comment
Repost
Share
Related threads
developer.joy's profile picture
Hypervel is a Laravel-style framework with native coroutine support for ultra-high performance.
テキストのグラフィックのようです
Like
8
Comment
1
Repost
Share
jandotai's profile picture
Skywork R1V2 (38B) now tops 70B+ multimodal systems on major reasoning benchmarks. It's closing the performance gap on OlympiadBench, AIME ’24, LiveCodeBench and MMMU. ​
カレンダー、、「R1V-2-38B 100 R1V-1-38B InternVL3-38B Qwen2.5VL-72B 72B Qwen2. Multimodal Reasoning Benchmarks R1V2-38B vs Open Source Models 80 QuQ-Preview-72B 72B 73.6 70.170.270.3 68 74 74.8 72.3 71.4 RN 67 Sennn 1g 60 40 62.6 49 20 40.4 52 51.1 ----------------. 33.2 34.2 38.135.9 5.9 MMMU(val) MathVista(mini) OlympiadBench MathVision(mini) mmmu-pro」というテキストの画像のようです
Like
5
Comment
2
Repost
Share
yannlecun's profile picture
Had to be said.
This is the 🔥
Audio is muted
Instagram logoguardian_us
Like
364
Comment
6
Repost
32
Share
5
theturingpost's profile picture
The freshest AI/ML research of the weekOur top 7:▪️ Syzygy of Thoughts ▪️ Sleep-time Compute ▪️ Retrieval-Augmented Generation with Conflicting Evidence ▪️ NodeRAG ▪️ DataDecide ▪️ ReTool ▪️ Antidistillation Sampling▪️ Thought Manipulation ▪️ Heimdall: Test-time scaling on the generative verification ▪️ Could Thinking Multilingually Empower LLM Reasoning? ▪️ Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?🧵
テキストの画像のようです
青写真・設計図、、「Rollout - Policy LLM Reward (a) Text-based RL Ltraining Question Advantage Rollout with Interleaved Code Execution Text: Policy LLM Code: Advantage Interpreter Feedback: Final Result: Code Sandbox Reward (b) ReTool RL training」というテキストの画像のようです
青写真・設計図、間取り図、、「IModule Task Problem bexi launched 3.Freeness Profite Generate necessary initial problem. raTp P(0.5 arop conditiones frow Spring( 6.5N/. friction 0.3m) 5.Exactness Horizontal lauch Logieal Soumdness velmarion 2.Betti Number Hleights (table Mapping(MI, M2, m3 Traverse ramp Mn) Task 5m) Dimersions Spring releases Cn. Mass Si Profite LLM 0.06kg) energy ออ่น №иeT M2 Inmediately flies 1....,C:-1} fi(module, fori>l for Frecness M3 Lands. Strategy 4.Mapping Freeeas Purpare Strategy- Resule CI. C2. Step-linked Strafegy-M2 CTCTpP gaps 0.2925 C3. Defined conditions Convert velocity Strafegr-M3 C3. 3.5m/s Evaluation fall heighr 6.Minimaliry 0.84n Strateg-M redundancy C2. Accupacy 日v・t=2.94日 c Complexity Simplicity simuleaneous 2.94 Paftern Matching Aggregate」というテキストのグラフィックのようです
Like
20
Comment
11
Repost
2
Share
2
Log in to see more replies.
Log in or sign up for ThreadsSee what people are talking about and join the conversation.
Instagram
Continue with Instagram
Log in with username instead