FAPO: Flawed-Aware Policy Optimization for Efficient and Reliable Reasoning. Project Page: https://fapo-rl.github.io/
Ding
dyyyyyyyy
·
AI & ML interests
None yet
Recent Activity
updated a dataset 7 days ago
dyyyyyyyy/mimo-v26-rl-oss-code-terminal-harbor published a dataset 12 days ago
dyyyyyyyy/mimo-v26-rl-oss-code-terminal-harbor updated a dataset 4 months ago
dyyyyyyyy/swe-rebench-filtered-1150Organizations
ScaleQuest
We introduce ScaleQuest, a scalable and novel data synthesis method. Project Page: https://scalequest.github.io/
-
Unleashing Reasoning Capability of LLMs via Scalable Question Synthesis from Scratch
Paper • 2410.18693 • Published • 41 -
dyyyyyyyy/ScaleQuest-Math
Viewer • Updated • 1M • 310 • 24 -
dyyyyyyyy/ScaleQuest-Code
Viewer • Updated • 157k • 60 • 4 -
dyyyyyyyy/ScaleQuest-Math-Qwen2.5
Viewer • Updated • 622k • 44
COLDQA
SCAN
We propose Self-Denoising Monte Carlo Annotation (SCAN), an efficient Process Reward Model (PRM) data synthesis and noise-tolerant learning framework.
GNER
We introduce GNER, a Generative Named Entity Recognition framework, which demonstrates enhanced zero-shot capabilities across unseen entity domains.
-
Rethinking Negative Instances for Generative Named Entity Recognition
Paper • 2402.16602 • Published • 3 -
dyyyyyyyy/GNER-LLaMA-7B
Text Generation • 7B • Updated • 71 • 4 -
dyyyyyyyy/GNER-T5-base
Text Generation • 0.2B • Updated • 127 • 2 -
dyyyyyyyy/GNER-T5-large
Text Generation • 0.8B • Updated • 22 • 2
FAPO
FAPO: Flawed-Aware Policy Optimization for Efficient and Reliable Reasoning. Project Page: https://fapo-rl.github.io/
SCAN
We propose Self-Denoising Monte Carlo Annotation (SCAN), an efficient Process Reward Model (PRM) data synthesis and noise-tolerant learning framework.
ScaleQuest
We introduce ScaleQuest, a scalable and novel data synthesis method. Project Page: https://scalequest.github.io/
-
Unleashing Reasoning Capability of LLMs via Scalable Question Synthesis from Scratch
Paper • 2410.18693 • Published • 41 -
dyyyyyyyy/ScaleQuest-Math
Viewer • Updated • 1M • 310 • 24 -
dyyyyyyyy/ScaleQuest-Code
Viewer • Updated • 157k • 60 • 4 -
dyyyyyyyy/ScaleQuest-Math-Qwen2.5
Viewer • Updated • 622k • 44
GNER
We introduce GNER, a Generative Named Entity Recognition framework, which demonstrates enhanced zero-shot capabilities across unseen entity domains.
-
Rethinking Negative Instances for Generative Named Entity Recognition
Paper • 2402.16602 • Published • 3 -
dyyyyyyyy/GNER-LLaMA-7B
Text Generation • 7B • Updated • 71 • 4 -
dyyyyyyyy/GNER-T5-base
Text Generation • 0.2B • Updated • 127 • 2 -
dyyyyyyyy/GNER-T5-large
Text Generation • 0.8B • Updated • 22 • 2
COLDQA