Aamer Mihaysi
O96a
AI & ML interests
Ethical AI, NLP & Cognitive architectures
Recent Activity
commentedon a paper about 4 hours ago
Precision on Demand: Escalating Weight Bits at Uncertain Decoding Steps in a Small Local Model commentedon a paper 1 day ago
Don't Just Trust Your Memory: Evidence-Linked Memory Reduces Memory-Induced Degradation in LLM Agents commentedon a paper 2 days ago
How Much Training-Inference Mismatch Can On-Policy Distillation Tolerate?Organizations
spaces 56
Build error
Agents
MMSkills Demo - Multimodal Skill Retrieval for Visual Agents
🤖
Build error
Agents
MMSkills Demo - Multimodal Skill Retrieval for Visual Agents
🤖
Sleeping
Agents
FEST-Style Few-Shot RL for Reasoning
🧠
Solve math problems with step‑by‑step reasoning
Runtime error
Agents
FEST-Style Few-Shot RL for Reasoning
🧠
Runtime error
Agents
Exp-007: Multi-Agent Pathfinding with Communication
🤖
Sleeping
Agents
Implicit Memory Conflict Validator
🧠
Evaluate LLM responses for outdated memory conflicts