LinkedIn Ships LLM-Powered Search Ranking at Scale

Also Worth Noting

• Steer2Adapt: Dynamically Composing Steering Vectors Elicits Efficient Adaptation of LLMs score 2
  关键词(5): efficient, lightweight, inference, reasoning, safety
• Cross-View World Models score 2
  关键词(3): agent, agents, planning
• VertCoHiRF: Decentralized Vertical Clustering Beyond k-means score 2
  关键词(2): agent, agents
• Fin-RATE: A Real-world Financial Analytics and Tracking Evaluation Benchmark for LLMs on SEC Filings score 2
  关键词(6): deployment, retrieval-augmented, reasoning, benchmark, evaluation
• Progressive Searching for Retrieval in RAG score 2
  关键词(4): efficient, RAG, search, cost
• Principled Synthetic Data Enables the First Scaling Laws for LLMs in Recommendation score 2
  关键词(5): scaling, preference, pre-training, recommendation, synthetic data
• KRONE: Hierarchical and Modular Log Anomaly Detection score 2
  关键词(2): efficient, efficiency
• Parallel Track Transformers: Enabling Fast GPU Inference with Reduced Synchronization score 2
  关键词(8): efficient, efficiency, fast, serving, inference
• Semantic Search At LinkedIn score 2
  关键词(10): scaling, efficiency, compression, distillation, pruning
• LUCID-SAE: Learning Unified Vision-Language Sparse Codes for Interpretable Concept Discovery score 2
  关键词(4): alignment, multimodal, vision-language, evaluation
• Beyond Accuracy: Risk-Sensitive Evaluation of Hallucinated Medical Advice score 2
  关键词(2): evaluation, safety
• Action-to-Action Flow Matching score 2
  关键词(7): efficiency, fast, inference, latency, real-time
• High Fidelity Textual User Representation over Heterogeneous Sources via Reinforcement Learning score 2
  关键词(2): latency, search
• Intent Mismatch Causes LLMs to Get Lost in Multi-Turn Conversation score 2
  关键词(2): scaling, alignment
• RAPiD: Real-time Deterministic Trajectory Planning via Diffusion Behavior Priors for Safe and Efficient Autonomous Driving score 2
  关键词(8): efficient, deployment, real-time, diffusion, planning