Channel Selection for Mixture-of-Experts Decoding Needs No Calibration Data
Paper • 2610.37386 • Published • 23 •
6
Guilherme Monteiro
guicybercode
AI & ML interests
smol-llms, llm-training, chinese-history, brazilian-history, production-ml, terminal-ui, data-curation, edge-ai
Recent Activity
commentedon a paper about 15 hours ago
Channel Selection for Mixture-of-Experts Decoding Needs No Calibration Data commentedon a paper about 15 hours ago
Extreme Compression of Per-Layer Embeddings via Cross-Layer Latent Vector Quantization commentedon a paper 3 days ago
Agent Mosaic: Recognizing Confidentiality Leaks in Agentic LLM SystemsOrganizations
commented 2
papers about 15 hours ago
commented 3
papers 3 days ago
Agent Mosaic: Recognizing Confidentiality Leaks in Agentic LLM Systems
Paper • 2610.30418 • Published • 25 •
11
Trust Mechanisms for Offline Generative AI at the Edge: A Connectivity-Regime Framework and an Empirical Study on a Raspberry Pi 5
Paper • 2610.33732 • Published • 5 •
13
Constrained Decoding: Harmful or Beneficial? Revisiting Its Impact on LLM Performance
Paper • 2610.35328 • Published • 8 •
6
CPU-only latency and a GGUF build of the adapter
#1 opened 3 days ago
by guicybercode
commented 3
papers 3 days ago
Constrained Decoding: Harmful or Beneficial? Revisiting Its Impact on LLM Performance
Paper • 2610.35328 • Published • 8 •
6
Channel Selection for Mixture-of-Experts Decoding Needs No Calibration Data
Paper • 2610.37386 • Published • 23 •
6
ThinkBit: Search Codewords, Recalibrate Scales, Select by Reasoning
Paper • 2610.34886 • Published • 3 •
3
commented 3
papers 4 days ago
Beyond Benchmarks: Can Small Open-Weight Models Get Real Coding Work Done?
Paper • 2610.34241 • Published • 10 •
7
Trust Mechanisms for Offline Generative AI at the Edge: A Connectivity-Regime Framework and an Empirical Study on a Raspberry Pi 5
Paper • 2610.33732 • Published • 5 •
13
Different Tasks, Different Timesteps: Temporal Task-Aware Quantization for Diffusion Language Models
Paper • 2610.31198 • Published • 24 •
5
updated a
Space 5 days ago
published a
Space 5 days ago
What should we compare next?
3
#3 opened 23 days ago
by 
pbhappliedsystems
commented 3
papers 5 days ago
Extreme Compression of Per-Layer Embeddings via Cross-Layer Latent Vector Quantization
Paper • 2610.31309 • Published • 18 •
7
Beyond Benchmarks: Can Small Open-Weight Models Get Real Coding Work Done?
Paper • 2610.34241 • Published • 10 •
7
Trust Mechanisms for Offline Generative AI at the Edge: A Connectivity-Regime Framework and an Empirical Study on a Raspberry Pi 5
Paper • 2610.33732 • Published • 5 •
13
Which of the nine changes carry over to AVX2-only CPUs?
#1 opened 5 days ago
by guicybercode
commented a
paper 5 days ago