Entdecken

30 technologies shown

• Neural Network concept Layered composition of weighted sums and nonlinearities, trained by gradient descent.
• Overfitting concept A model fitting its training data instead of the pattern behind it.
• Prompt Caching technique Reusing the computed state of an unchanged prompt prefix.
• Prompt Engineering technique Shaping a model's instructions and context to get the output you actually want.
• RoPE technique Rotary position embedding, encoding position as a rotation of the query and key.
• SQuAD dataset Questions written by people about Wikipedia articles with answer spans.
• Speculative Decoding technique Generating several tokens with a fast model and verifying them with the real one.
• Tensor Core hardware Dedicated matrix-multiply units inside NVIDIA GPUs.
• Tokenizer tool Component that converts text to tokens and back.
• Top-p Sampling technique Sampling only from the smallest set of tokens whose probabilities reach p.
• V100 hardware The first NVIDIA GPU with tensor cores, and the one early transformers trained on.
• Alignment technique Making a model's behaviour match what people want, not just what is likely.
• B200 hardware NVIDIA's Blackwell-generation datacentre GPU.
• Context Engineering technique Deciding what goes into the context window, not just how it is worded.
• Few-Shot Learning technique Putting examples in the prompt instead of training on them.
• QLoRA technique Fine-tuning a quantised model with LoRA adapters, so a large model fits on one GPU.
• Beam Search technique Keeping several candidate sequences alive and extending the best.
• Cohere Embed model Hosted multilingual embedding model with separate query and document modes.
• Diffusers library Library of diffusion model implementations and pipelines.
• Dropout technique Randomly zeroing activations during training to reduce overfitting.
• FAISS library Library for efficient similarity search over dense vectors.
• Feed-Forward Network concept The position-wise fully connected layer inside each transformer block.
• GRPO technique Policy optimisation that scores a group of answers relative to each other.
• GSM8K benchmark Grade-school mathematics word problems requiring multi-step reasoning.
• Gemini model Google's hosted multimodal model family, integrated with its cloud platform.
• GraphRAG technique Retrieval over a knowledge graph built from the corpus.
• LM Evaluation Harness tool Standard harness for running language model benchmarks.
• Machine Learning concept Field of building systems that improve at a task from data rather than from explicit rules.
• Multi-Agent technique Several agents working on a task, with defined roles or handoffs.
• PEFT library Parameter-efficient fine-tuning methods behind one interface.