Browser-tab grouping with a 0.8B model: LoRA adapters on Qwen3.5-0.8B (SFT, GRPO) and the synthetic training data.
👋 Open to Work
Manuel Parra
th3nolo
AI & ML interests
Based in Miami, FL.
Recent Activity
liked a model 12 days ago
baselabs/joey-1 updated a collection 21 days ago
TabCrew updated a collection 21 days ago
TabCrewOrganizations
None yet
-
openai/gpt-oss-120b
Text Generation • 117B • Updated • 2.44M • • 5.51k -
Emergent temporal abstractions in autoregressive models enable hierarchical reinforcement learning
Paper • 2512.20605 • Published • 62 -
Nested Browser-Use Learning for Agentic Information Seeking
Paper • 2512.23647 • Published • 19 -
TimeBill: Time-Budgeted Inference for Large Language Models
Paper • 2512.21859 • Published • 25
Browser-tab grouping with a 0.8B model: LoRA adapters on Qwen3.5-0.8B (SFT, GRPO) and the synthetic training data.
-
openai/gpt-oss-120b
Text Generation • 117B • Updated • 2.44M • • 5.51k -
Emergent temporal abstractions in autoregressive models enable hierarchical reinforcement learning
Paper • 2512.20605 • Published • 62 -
Nested Browser-Use Learning for Agentic Information Seeking
Paper • 2512.23647 • Published • 19 -
TimeBill: Time-Budgeted Inference for Large Language Models
Paper • 2512.21859 • Published • 25