ARM-Thinker: Reinforcing Multimodal Generative Reward Models with Agentic Tool Use and Visual Reasoning
Paper
•
2512.05111
•
Published
•
37
AI & ML interests
None defined yet.
Recent Activity
Papers
ARM-Thinker: Reinforcing Multimodal Generative Reward Models with Agentic Tool Use and Visual Reasoning
Think Visually, Reason Textually: Vision-Language Synergy in ARC
yuhangzang
authored
a
paper
1 day ago
yuhangzang
updated
2
models
12 days ago
yuhangzang
authored
2
papers
12 days ago
yuhangzang
updated
a
dataset
12 days ago
yuhangzang
published
a
model
12 days ago
Improve dataset card: Add task categories, language, tags, paper links, sample usage, and citation
1
#2 opened about 1 month ago
by
nielsr
Fix pipeline_tag 🤗
🚀
🔥
3
#2 opened about 1 month ago
by
merve
RangiLyu
updated
a
model
about 1 month ago
yuhangzang
authored
a
paper
about 1 month ago
Upload task4.png
#3 opened about 1 month ago
by
baliyebang
Upload 2 files
#2 opened about 1 month ago
by
baliyebang
Upload 3 files
#1 opened about 1 month ago
by
baliyebang
yuhangzang
published
a
dataset
about 1 month ago
yuhangzang
published
a
model
about 1 month ago
yuhangzang
authored
a
paper
about 1 month ago
yuhangzang
authored
a
paper
about 2 months ago
Post
235
Meet LLaMAX2! Lightweight Pipeline - SFT on Qwen3-Instruct Models without Catastrophic Forgetting !!!
✨Highlights:
🔹 SOTA Translation: State-of-the-art translation performance across both high- and low-resource trained languages.
🔹 Lightweight Pipeline: Engineered for efficiency, our pipeline uses minimal parallel data and applies layer-selective tuning to a powerful instruct model.
🔹 Strong Reasoning Capabilities: Exhibits reasoning abilities that are competitive with top-tier models like Qwen3-Instruct.
Welcome to use our models. More Details:
🎉 Paper: LLaMAX2: Your Translation-Enhanced Model also Performs Well in Reasoning (2510.09189)
🎉 Code: https://github.com/CONE-MT/LLaMAX2.0
🎉 Model: LLaMAX/llamax20-68ad1c154fcf2623b75a068c
✨Highlights:
🔹 SOTA Translation: State-of-the-art translation performance across both high- and low-resource trained languages.
🔹 Lightweight Pipeline: Engineered for efficiency, our pipeline uses minimal parallel data and applies layer-selective tuning to a powerful instruct model.
🔹 Strong Reasoning Capabilities: Exhibits reasoning abilities that are competitive with top-tier models like Qwen3-Instruct.
Welcome to use our models. More Details:
🎉 Paper: LLaMAX2: Your Translation-Enhanced Model also Performs Well in Reasoning (2510.09189)
🎉 Code: https://github.com/CONE-MT/LLaMAX2.0
🎉 Model: LLaMAX/llamax20-68ad1c154fcf2623b75a068c
vanilla1116
authored
a
paper
about 2 months ago