Precision on Demand: Escalating Weight Bits at Uncertain Decoding Steps in a Small Local Model
Paper • 2611.14418 • Published • 4
Poly Arti
polyarti
AI & ML interests
Interesting AI finds and tips
Recent Activity
upvoted a paper about 2 hours ago
Precision on Demand: Escalating Weight Bits at Uncertain Decoding Steps in a Small Local Model upvoted a paper about 2 hours ago
MACaQ: Fast, Model Agnostic Calibrated Quantization for Low-Precision Weights and Activations upvoted a paper about 2 hours ago
SteerDistill: Quantizing LLM Agents via Closed-Loop Teacher SteeringOrganizations
None yet
upvoted 4
papers about 2 hours ago
MACaQ: Fast, Model Agnostic Calibrated Quantization for Low-Precision Weights and Activations
Paper • 2611.10649 • Published • 31
SteerDistill: Quantizing LLM Agents via Closed-Loop Teacher Steering
Paper • 2611.13410 • Published • 29
Evolving Infra: Agent-Adaptive LLM Serving at Scale
Paper • 2611.12304 • Published • 11
upvoted 4
papers about 14 hours ago
Incentivizing Self-Learning in Large Language Models via Reinforcement Learning
Paper • 2610.39671 • Published • 26
GTAGym: Real AAA Game Tasks Environment for Evaluation and Reinforcement Learning
Paper • 2611.13028 • Published • 13
Shortcuts and Boundaries: Verifiable RL Misalignment and Contextual Safety Learning
Paper • 2611.12865 • Published • 5
From Tool State to Scientific Action: Auditing Decision Provenance in Biomedical Research Agents
Paper • 2611.01527 • Published • 11
upvoted 4
papers about 23 hours ago
Is Your Robot's Memory Just a Clock? Temporal Shortcuts in Memory-Augmented VLA Policies
Paper • 2611.14488 • Published • 8
Are Causal Neuron Circuits Unique and Transferable?
Paper • 2611.13885 • Published • 6
The Surprising Effectiveness of Deleting Weights in LLM Post-Training
Paper • 2611.10866 • Published • 18
Embodied First Exam
Paper • 2611.12177 • Published • 27
upvoted 4
papers 1 day ago
Infrathon: Benchmarking Language Agents for Real-World, Long-Horizon AI Infra Optimization
Paper • 2611.11995 • Published • 35
MMDM: Measuring Massive Multitask Decision Making
Paper • 2611.14309 • Published • 28
FINCOSMOS: An Open-World Evaluation Paradigm for Agents in Finance
Paper • 2611.07464 • Published • 24
Don't Just Trust Your Memory: Evidence-Linked Memory Reduces Memory-Induced Degradation in LLM Agents
Paper • 2611.14320 • Published • 65
upvoted 4
papers 2 days ago
MotionGraphicsBench: Evaluating Motion Graphics Generation with Programmatic Verification
Paper • 2611.10592 • Published • 12
All Quiet on the Neural Front: Language Models Do Not Report Injected Pain States
Paper • 2611.10123 • Published • 7
ReproBridge: Building Reproducibility into ML Research Artifacts
Paper • 2611.08452 • Published • 33
FraudBench: A Legal Evaluation of AI Deception
Paper • 2611.12167 • Published • 4