🧰 Llama

The items that name it 37

• Does Chain-of-Thought Reliance Follow a Scaling Law?Larger Models Follow Their Steps More, Independent of Post-Training
  paperas classification.technologies2026-10-28
• FlatCake: A Flat KVCache Storage Architecture for LLM Serving
  paperas classification.technologies2026-10-28
• Heads of Security: Learned Attention Gates Improve Prompt Injection Robustness
  paperas classification.technologies2026-10-27
• Folding Attention Biases into Keys Makes Compacted KV Caches Run on Standard Kernels
  paperas classification.technologies2026-10-27
• What Adobe v ArtCraft teaches us
  articleas models2026-10-25
• MPQ: A Message-Passing View of Post-Training Quantization
  paperas models2026-10-23
• Training Small Language Models for Transferable Tool Selection
  paperas classification.technologies2026-10-22
• How Specializing the Vocabulary of SLMs Can Cut Inference Cost by 77% in Real Applications
  articleas classification.technologies2026-10-21
• FairStress: A deeper evaluation of fairness in LLM decision-making
  articleas models2026-10-20
• HERMES: High-Performance Private Decoding for Large Language Models
  paperas classification.technologies2026-10-18
• Meta shakes up the frontiers, open and closed
  articleas classification.technologies2026-10-17
• Rodrigo ES Porto (@RodrigoESPorto) on X
  threadas classification.technologies2026-10-16
• FuTransformer: FP8 prefill and INT4 decode for Llama-style inference on Hopper
  paperas classification.technologies2026-10-16
• Multi-Reward Hacking: From Aggregation-Induced Drift to Trust-Anchored Defense
  paperas models2026-10-15
• MultiAlloc: Budget-Aware Precision Allocation for KV Cache Quantization
  paperas models2026-10-13
• Geometric Calibration of Supervised Fine-Tuning
  paperas classification.technologies2026-10-13
• Meta's New Muse Wonder Finally Puts Zuckerberg in the Frontier AI Race
  articleas classification.key entities2026-10-12
• Anthropic tests AI autonomous replication across eight nodes in four of five trials
  articleas classification.technologies2026-10-12
• AI models may be capable of self-replicating without access to a datacenter
  writeupas classification.technologies2026-10-12
• DualQuant: Dual-scale Calibration-free Weight-and-Activation Quantization for LLMs
  paperas models2026-10-09
• Sub-1-Bit LLM Compression via Latent Factorization
  repoas models2026-10-08
• SLDR: Defending Against Malicious Fine-tuning via Selective Layers Recovery and Dynamic Routing
  paperas classification.technologies2026-10-07
• Principled Under Pressure: Post-Training Decides Whether LLMs Act on Their Own Moral Judgment
  paperas classification.technologies2026-10-06
• Defense-in-Depth for LLMs: Evaluating Memory Gates Against Activation-Induced and Memory-Induced Sycophancy
  paperas classification.technologies2026-10-05
• Sujeet (@sujeetgholap) on X
  threadas models2026-10-03
• Inside-Out AI: Rebuilding Airbnb Behind the Scenes and Across the Guest Experience
  articleas models2026-10-02
• Identity Management for Agentic AI [pdf] (2025)
  paperas models2026-10-01
• Uncensored and Offensive Security AI Models Benchmark
  repoas classification.key entities2026-09-29
• OpenRouter: from Seed to Stripe — with OpenRouter’s Alex Atallah & AMP’s Anjney Midha
  talkas models2026-09-25
• ClusterMAX 3.0: The Industry Standard GPU Cloud Rating System Returns
  articleas classification.key entities2026-09-23
• The current balance of power in open models
  articleas models2026-09-21
• A Zeroth-Order Paradigm for LLM Preference Alignment
  paperas models2026-09-16
• Long Live the Short King: Why 4-hi HBM Wins
  articleas classification.key entities2026-09-13
• Latest open artifacts (#24): Motif-3, GLM-5.3, Hy4-preview and open model licenses
  articleas models2026-09-08
• TPU Inference Externalization Full Steam Ahead - InferenceX
  articleas classification.key entities2026-09-07
• rikugan: LLM reverse-engineering agent for IDA Pro and Binary Ninja
  repoas classification.key entities
• LLM Inference in a Flash!
  paperas models

In every workspace you can read 54