🎯 AI: techniques

Attack and defence techniques and machine learning methods that a second item could name too, with an ATT&CK, ATLAS or CWE id when one is stated. A procedure only one article describes is item text. Each one opens the items that name it, and the list is ranked by how many do.

This workspace's kept items name 295 of them. This is page 3 of 3, the most named first.

• Mid-training on multimodal corpus2 items
• Mid-turn steering2 items
• Model Context Protocol2 items
• Model distillation2 items
• Model organisms2 items
• Multi-agent orchestration2 items
• Multi-domain on-policy distillation2 items
• Multi-Head Latent Attention2 items
• Multi-layer defense against reward hacking2 items
• Multi-Prefix Multi-Teacher On-Policy Distillation2 items
• Multi-robot collaboration2 items
• Muon2 items
• noul questions2 items
• NVFP4 expert-only quantization2 items
• Obligation binding2 items
• On-policy self distillation2 items
• Paged KV cache2 items
• Parallel box decoding2 items
• Partial RoPE2 items
• Pass@k2 items
• Per-agent git worktrees2 items
• Physical settlement of compute2 items
• Post-training2 items
• Posture as a coordinate per phase2 items
• PPO2 items
• Preparedness Framework2 items
• Promptbreeder2 items
• QLoRA2 items
• RadixAttention2 items
• Reachability analysis2 items
• ReAct2 items
• Reasoning effort control2 items
• Reasoning effort scaling2 items
• Reasoning models2 items
• Reasoning summarization2 items
• recall2 items
• Recall-then-diverge trajectory grading2 items
• Recommendation system over a connected graph2 items
• Recursive Language Models2 items
• Recursive self-improvement of cache policies2 items
• Reflexion2 items
• Refusal-adjusted scoring2 items
• Reinforcement Learning for Calibrated Decisions2 items
• ReplaySSM2 items
• Repo-based project record2 items
• repos2 items
• Representation Autoencoders2 items
• Reward hacking2 items
• RL for Calibrated Decisions2 items
• RLCR2 items
• Router cascade2 items
• Sandwich norm2 items
• Scheduled deletion2 items
• Scheduled Tasks2 items
• Score centering2 items
• Selection-based Structured Reasoning2 items
• Self-Consistency2 items
• Self-feedback RL2 items
• Self-hosting on batch-invariant kernels2 items
• Self-Refine2 items
• Self-speculative decoding2 items
• Semantic Caching2 items
• Separating the brain from the hands2 items
• Serverless GPU fast startup2 items
• SFT2 items
• Speaker diarization2 items
• Stretchtext2 items
• Structured outputs2 items
• SynthID-Text2 items
• System One decision model2 items
• Task refinement2 items
• Teacher forcing2 items
• Ternary quantization2 items
• Test-Driven Development2 items
• Test-Time Training2 items
• Text watermarking2 items
• TF-IDF2 items
• Thinking mode2 items
• Token-gated attention2 items
• Tool-response rolling-drop cache policy2 items
• Tri-hybrid layer mix2 items
• Truncated backpropagation2 items
• Truncated importance sampling2 items
• Two-class amendment2 items
• Two-sided admission2 items
• Video frame sampling2 items
• Watermarking2 items
• Wide expert parallelism2 items
• Wide expert parallelism with expert replication2 items
• Workload Identity Federation2 items
• World model2 items
• YaRN RoPE scaling2 items
• You Only RL Once2 items
• Hybrid retrieval1 item
• RAG Pipeline1 item