Rethinking the Multilingual Reasoning Gap with Layer Swap
Paper • 2605.26735 • Published • 2
KuraKura AI
community
AI & ML interests
Sea Turtles
Recent Activity
View all activity
MaxLSB
authored a
paper 4 days ago
Post
2212
Big update to llm-datasets, my curated list of datasets and tools for post-training LLMs.
> Added many new datasets
> New "thinking" column
> Refreshed recommended tools.
Thanks to everyone who told me they used it for their research at ICLR, you motivated this update!
> Added many new datasets
> New "thinking" column
> Refreshed recommended tools.
Thanks to everyone who told me they used it for their research at ICLR, you motivated this update!
- 2 replies
adds Q4_KM for mobile compatible
#1 opened 3 months ago
by 
Tonic
adds Q4 KM gguf file for mobile compatibility
1
#1 opened 3 months ago
by Tonic
mlabonne
authored 2
papers 4 months ago
Post
10389
New family of 1B models just dropped!
> LiquidAI/LFM2.5-1.2B-Base: 10T → 28T tokens
> LiquidAI/LFM2.5-1.2B-Instruct: new large-scale multi-stage RL
> LiquidAI/LFM2.5-1.2B-JP: our most polite model
> LiquidAI/LFM2.5-VL-1.6B: multi-image multilingual
> LiquidAI/LFM2.5-Audio-1.5B: 8x times faster, no quality loss
Super proud of this release 🤗
> LiquidAI/LFM2.5-1.2B-Base: 10T → 28T tokens
> LiquidAI/LFM2.5-1.2B-Instruct: new large-scale multi-stage RL
> LiquidAI/LFM2.5-1.2B-JP: our most polite model
> LiquidAI/LFM2.5-VL-1.6B: multi-image multilingual
> LiquidAI/LFM2.5-Audio-1.5B: 8x times faster, no quality loss
Super proud of this release 🤗
- 3 replies
MaxLSB
authored a
paper 8 months ago
GAD-cell
updated 2
datasets 8 months ago
GAD-cell
updated 5
models 8 months ago
kurakurai/Luth-0.6B-Instruct
Text Generation • 0.6B • Updated • 1.32k • • 9
kurakurai/Luth-1.7B-Instruct
Text Generation • 2B • Updated • 111 • • 14
kurakurai/Luth-LFM2-350M
Text Generation • 0.4B • Updated • 258 • 15
kurakurai/Luth-LFM2-700M
Text Generation • 0.7B • Updated • 141 • 16
kurakurai/Luth-LFM2-1.2B
Text Generation • 1B • Updated • 165 • 26
GAD-cell
authored a
paper 8 months ago
Post
8463
LiquidAI/LFM2-8B-A1B just dropped!
8.3B params with only 1.5B active/token 🚀
> Quality ≈ 3–4B dense, yet faster than Qwen3-1.7B
> MoE designed to run on phones/laptops (llama.cpp / vLLM)
> Pre-trained on 12T tokens → strong math/code/IF
8.3B params with only 1.5B active/token 🚀
> Quality ≈ 3–4B dense, yet faster than Qwen3-1.7B
> MoE designed to run on phones/laptops (llama.cpp / vLLM)
> Pre-trained on 12T tokens → strong math/code/IF
- 1 reply
Post
3910
⚛️ New drop of tiny task-specific models!
Want to do data extraction, translation, RAG, tool use, or math on a Raspberry Pi? We got you covered! ✅
These tiny models were fine-tuned to perform narrow tasks extremely well, making them competitive with much larger models.
You can deploy them today on-device or even on GPUs for big data operations!
LiquidAI/liquid-nanos-68b98d898414dd94d4d5f99a
Want to do data extraction, translation, RAG, tool use, or math on a Raspberry Pi? We got you covered! ✅
These tiny models were fine-tuned to perform narrow tasks extremely well, making them competitive with much larger models.
You can deploy them today on-device or even on GPUs for big data operations!
LiquidAI/liquid-nanos-68b98d898414dd94d4d5f99a
- 1 reply
MaxLSB
updated 2
models 8 months ago