LLM Agents
ai agents, autonomous agents, reasoning
An AI agent is a system built around a large language model that does not just answer a question but pursues a goal by taking actions in a loop. Where a plain chatbot maps one…
27,049 technical terms and definitions
ai agents, autonomous agents, reasoning
An AI agent is a system built around a large language model that does not just answer a question but pursues a goal by taking actions in a loop. Where a plain chatbot maps one…
rag, agents, architecture, building ai, langchain, llamaindex, production systems
Building LLM applications involves architecting systems that integrate language models with data, tools, and user interfaces — choosing appropriate patterns like RAG or agents…
auto eval, gpt4
LLM As Judge LLM-as-judge uses a strong language model to evaluate outputs from weaker models or different systems providing scalable automated evaluation. GPT-4 commonly serves…
evaluation
LLM-as-Judge is an evaluation paradigm where a strong language model (typically GPT-4 or Claude) is used to evaluate the quality of outputs from other models, replacing…
beginner, tokens, prompts, context window, temperature, getting started, ai fundamentals
LLM basics for beginners provides a foundational understanding of how large language models work and how to use them effectively — explaining core concepts like tokens, prompts…
mmlu, hellaswag, gsm8k, human eval, lm evaluation harness
Evaluating a large language model is harder than evaluating almost any software that came before it, because the thing you want to measure — general competence and good behavior…
github copilot, codex code llm, code completion neural, deepseekcoder code model
LLM Code Generation: From Codex to DeepSeek-Coder — transformer models for code completion and synthesis Code generation via large language models (LLMs) has transformed developer…
LLM-Powered Conversational Recommender Agents Introduction & Motivation LLM-Powered Conversational Recommender Agents addresses a central problem in recsys systems and {target}…
llm evals, evals, llm behavior, evaluating llms, how to evaluate llms, llm evaluation benchmark, model evaluation metrics, llm as a judge
Evaluating a large language model is harder than evaluating almost any software that came before it, because the thing you want to measure — general competence and good behavior…
llm evaluation, benchmarking contamination measuring, ci equal gsm8k helm irt llm, llm evaluation benchmarking & contamination measuring progress reliably
LLM Evaluation, Benchmarking & Contamination: Measuring Progress Reliably Introduction & Motivation As large language models have grown more capable, the task of measuring that…
llm hallucination, hallucination detection, ai api dpo frank llm nli, llm hallucination detection & mitigation, llm hallucination detection, hallucination detection mitigation, detection mitigation
LLM Hallucination Detection & Mitigation 1. Introduction & Motivation Large language models routinely produce fluent, confident, and grammatically flawless text that is…
grounded generation, retrieval augmented generation hallucination, factual consistency, faithfulness llm
LLM Hallucination Mitigation is the collection of techniques — architectural, training-time, and inference-time — designed to reduce the rate at which Large Language Models…
vllm pagedattention throughput tuning, tensorrt llm triton deployment pipeline, kv cache continuous batching, quantized inference gptq awq gguf
LLM Inference Serving Optimization Stack is the runtime layer that converts trained models into reliable, low-latency, cost-efficient production services. For most enterprises…
llm jailbreaking red teaming adversarial prompting defenses, llm jailbreaking, red teaming adversarial, ai api asr dan dpo gcg, llm jailbreaking red-teaming, adversarial prompting
LLM Jailbreaking, Red-Teaming & Adversarial Prompting Defenses 1. Introduction & Motivation Instruction-tuned and RLHF-aligned language models are trained to refuse harmful…
latency, throughput, quantization, kv cache, flash attention, speculative decoding, vllm, inference optimization
LLM optimization is the systematic process of improving inference speed, reducing latency, and maximizing throughput — using techniques like quantization, KV cache optimization…
posttraining fine tuning pipeline, sft supervised fine tuning llm, lora low rank adaptation llm, qlora quantized adapter tuning, peft adapter prefix prompt tuning, llm finetuning ab testing deployment
Post-training Fine-tuning Pipeline converts a generic base model into an instruction-following system tuned for target domains, policies, and user experience requirements.…
data curation llm, training data quality, web crawl filtering, common crawl, data mixture
LLM Pretraining Data Curation is the systematic process of collecting, filtering, deduplicating, and mixing text corpora to create the training dataset for large language models…
foundation model pretraining pipeline, distributed llm training parallelism, tokenizer bpe sentencepiece vocabulary, zero fsdp optimizer sharding
Pre-training LLM Foundation Models is the full-stack process of building a base model from raw text and code corpora through tokenizer design, architecture selection, distributed…
prompt injection llm attack, llm bias fairness, model collapse training, responsible ai deployment
LLM Safety and Responsible Deployment: Jailbreaking, Bias, and Scaling Policies — navigating safety risks at scale Large language models exhibit safety vulnerabilities…
ai generated text detection, watermark language model, green red token list, detecting ai text
LLM Watermarking and AI Text Detection is the technique of embedding imperceptible statistical signatures into AI-generated text during generation — allowing detection…
lmql, language model query language, framework
LMQL (Language Model Query Language) is a specialized programming language designed for interacting with large language models in a structured, controllable way. It combines…
local, gui
LM Studio is a desktop application for discovering, downloading, and running local LLMs through a polished graphical interface — providing a built-in Hugging Face Hub browser…
evaluation
LMSYS Chatbot Arena is the most prominent open platform for evaluating and comparing large language models through live human voting. Users submit prompts that are answered by two…
lna noise figure drift prediction, lna noise figure, noise figure drift, figure drift prediction, lna noise, noise figure, figure drift, drift prediction
LNA Noise Figure Drift Prediction with Machine Learning Introduction & Motivation LNA Noise Figure Drift Prediction with Machine Learning addresses a central problem in low-noise…
lna noise figure drift prediction, lna noise figure, noise figure drift, figure drift prediction, lna noise, noise figure, figure drift, drift prediction
LNA Noise Figure Drift Prediction with Machine Learning Introduction & Motivation LNA Noise Figure Drift Prediction with Machine Learning addresses a central problem in low-noise…
rf circuits modeling noise figure degradation in lnas due to process and environmental drift, rf circuits modeling, circuits modeling noise, modeling noise figure, noise figure degradation
RF Circuits: Modeling noise figure degradation in LNAs due to process and environmental drift Introduction & Motivation RF Circuits: Modeling noise figure degradation in LNAs due…
modeling noise figure degradation in lnas due to process and environmental drift, modeling noise figure, noise figure degradation, figure degradation lnas, degradation lnas due, lnas due
Modeling noise figure degradation in LNAs due to process and environmental drift Introduction & Motivation Modeling noise figure degradation in LNAs due to process…
nginx, reverse proxy
Load Balancing for ML Services Why Load Balance? Distribute traffic across multiple model instances for reliability, scalability, and efficient resource utilization. Load…
load balancer session affinity, load balancer session, balancer session affinity, load balancer, balancer session, session affinity
Load Balancer Session Affinity Introduction & Motivation Load Balancer Session Affinity addresses a central problem in cloud process control and session-persistence-cookie-based…
load balancer, l4 load balancing, l7 load balancing, consistent hashing, least connections, ai inference routing
Load balancing distributes work across healthy service instances to meet throughput, latency, locality and availability goals. AI inference needs routing that understands model…
manufacturing operations
Load Balancing is the distribution of work across equivalent tools or lines to avoid localized congestion - It is a core method in modern semiconductor operations execution…
ai agents
Load Balancing Agents is the distribution of workload across agents to prevent bottlenecks and idle capacity - It is a core method in modern semiconductor AI-agent coordination…
load balancing algorithm, load balancing, balancing algorithm
Load Balancing Algorithm Introduction & Motivation Load Balancing Algorithm addresses a central problem in cloud process control and distribution-fairness-equality prediction: how…
load balancing algorithm selection, load balancing algorithm, balancing algorithm selection, load balancing, balancing algorithm, algorithm selection
Load Balancing Algorithm Selection Introduction & Motivation Load Balancing Algorithm Selection addresses a central problem in distributed-systems process control…
operations
Load balancing dispatch is the dispatch strategy that distributes incoming lots across parallel tools to avoid queue concentration and uneven utilization - it improves flow…
moe
Load Balancing Loss is the auxiliary training objective added to Mixture of Experts models that penalizes uneven expert utilization — encouraging the router to distribute tokens…
architecture
Load Balancing Loss is auxiliary objective that encourages tokens to distribute more evenly across experts - It is a core method in modern semiconductor AI serving…
load balancing, moe, model architecture
Load balancing in MoE ensures experts are used roughly equally, preventing underutilization and bottlenecks. The problem: Without balancing, router may send most tokens to few…
work distribution, load imbalance
Load Balancing — distributing computational work evenly across parallel processors/threads so that no processor is idle while others are still working. The Problem Total parallel…
dynamic load balancing, work distribution, parallel load imbalance
Dynamic Load Balancing is the runtime distribution and redistribution of workload across parallel processing elements to minimize idle time and maximize throughput, addressing…
dynamic load balancing, work stealing, static load balance, parallel workload distribution
Load Balancing in Parallel Computing is the resource allocation discipline that distributes computational work evenly across available processors — preventing the scenario where…
dynamic load balance, work distribution, static dynamic scheduling, imbalanced workload parallel
Load Balancing in Parallel Computing is the algorithmic and runtime strategy for distributing work evenly across all processing elements — ensuring that no processor sits idle…
dynamic load balancing, static load balancing partitioning, work stealing load balance, load imbalance detection
Load Balancing in Parallel Computing is the process of distributing computational work evenly across all available processing units to minimize idle time and maximize throughput…
dynamic load balancing, static load partitioning, work distribution strategies, load imbalance overhead parallel
Load Balancing Strategies are techniques for distributing computational work across parallel processing elements to minimize idle time and maximize overall throughput — effective…
advanced test & probe
Load Board is a test hardware board that routes power and signals between ATE resources and packaged devices - It is optimized for signal integrity, thermal handling, and fixture…
automation
Load locks are vacuum-compatible chambers that transition wafers between atmospheric and vacuum environments. Purpose: Allow wafers to enter vacuum process chambers without…
manufacturing operations
Load Lock is an interface chamber that transfers wafers between atmospheric handling and vacuum process modules - It is a core method in modern semiconductor wafer handling…
load lock cycle time prediction, load lock cycle, lock cycle time, cycle time prediction, load lock, lock cycle, cycle time, time prediction
Load Lock Cycle Time Prediction with Machine Learning Introduction & Motivation Load Lock Cycle Time Prediction with Machine Learning addresses a central problem in vacuum load…
load lock door seal wear prediction machine learning, load lock, door seal wear, load lock door seal wear prediction, load lock door, lock door seal, seal wear prediction, lock door
Load Lock Door Seal Wear Prediction with Machine Learning Introduction & Motivation Load Lock Door Seal Wear Prediction with Machine Learning addresses a central problem in load…
load lock pump-down time drift prediction, load lock pump-down, lock pump-down time, pump-down time drift, time drift prediction, load lock, lock pump-down, pump-down time
Load Lock Pump-Down Time Drift Prediction with Machine Learning Introduction & Motivation Load Lock Pump-Down Time Drift Prediction with Machine Learning addresses a central…