Skip to content

Concept Glossary: L

This page lists English glossary entries for this letter. Entry bodies are assembled from term source files.

label

  • Meaning: A label is the target answer or target output attached to a data example for the current task. It defines what the model is supposed to predict, such as a class name or numeric target.
  • Why it matters: Labels define supervised-learning targets, so changing the label definition changes the task and the evaluation. Poor or ambiguous labels can limit model quality before training even begins.
  • Related concepts: example, data, output, labeling, target
  • Core Section: P1-4.2
  • Appears in: P1-2.1, P1-3.2, P1-8.1

labeling

  • Meaning: Labeling is the work of attaching target answers or category markers to data examples according to a documented rule.
  • Why it matters: Model quality can depend as much on labeling rules as on the algorithm. Labeling makes readers ask who defined the answer and by what standard.
  • Related concepts: label, supervised learning, target, task definition
  • Core Section: P1-8.1

language model

  • Meaning: A language model estimates how likely words or tokens are in sequence, often by predicting what can come next from previous context.
  • Why it matters: This concept places LLMs in the older lineage of probabilistic language prediction rather than treating them only as chat products.
  • Related concepts: language modeling, statistical language model, token
  • Core Section: P1-11.1
  • Appears in: P1-10.2, P1-11.2, P1-11.3, P6-19.1, P6-19.2

language modeling

  • Meaning: Language modeling is the task of estimating the probability structure of token or word sequences.
  • Why it matters: It explains why next-token prediction and sequence probability are central to modern LLMs and their earlier statistical and neural predecessors.
  • Related concepts: direct lineage, Transformer, embedding
  • Core Section: P1-9.3
  • Appears in: P1-10.1, P1-11.1, P1-11.3, P6-19.2

latency

  • Meaning: Latency is the time a user waits after sending a request until the first response or final result arrives.
  • Why it matters: Good model quality is not enough if users wait too long. Latency connects model, retrieval, tool calls, and postprocessing to service experience.
  • Related concepts: cost, throughput, streaming, retry
  • Core Section: P1-14.6
  • Appears in: P1-14.5

layer normalization

  • Meaning: Layer normalization normalizes values inside a representation for one case so later layers receive more stable value scales.
  • Why it matters: Deep Transformer blocks need stable value scales. Layer normalization helps repeated attention and feed-forward blocks remain trainable and usable.
  • Related concepts: Transformer, residual connection, feed-forward network
  • Core Section: P5-14.1
  • Appears in: P5-14.2

learned representation

  • Meaning: A learned representation is an internal form a model builds from data so useful differences for a task become easier to use.
  • Why it matters: The concept distinguishes hand-designed features from internal representations learned by a model for the current task.
  • Related concepts: representation, representation learning, hand-crafted features
  • Core Section: P1-9.1
  • Appears in: P5-10.1, P5-10.2

learning rate

  • Meaning: Learning rate controls how large each update step is when optimization moves in a direction that should reduce loss.
  • Why it matters: Learning rate controls the balance between stable progress and speed. Too large or too small a step can derail or slow training.
  • Related concepts: gradient descent, gradient, optimization
  • Core Section: P2-6.3
  • Appears in: P5-7.1, P5-7.2, P5-7.3

least privilege

  • Meaning: Least privilege is the security principle of giving a person, tool, or agent only the access required for the current task.
  • Why it matters: AI systems often connect to tools that change real state. Restricting access reduces the damage from mistakes, misuse, or prompt injection.
  • Related concepts: permission, approval, security, agent, tool use
  • Core Section: P1-15.3
  • Appears in: P1-14.5, P1-14.6, P7-6.2

legend

  • Meaning: A legend is the explanation box that tells what lines, colors, markers, or series in a chart represent.
  • Why it matters: Without a correct legend, chart comparisons can be misread. It links visual marks to data series and makes interpretation possible.
  • Related concepts: plot, accuracy, loss curve
  • Core Section: P2-13.3
  • Appears in: P2-15.1

limit

  • Meaning: A limit describes what value a function approaches as the input gets close to a point or condition.
  • Why it matters: Limits support the language of change, continuity, and derivatives, especially when direct substitution is not enough.
  • Related concepts: convergence, function, rate of change
  • Core Section: P2-2.3
  • Appears in: P2-4.1, P2-4.2

limit of prompting

  • Meaning: The limit of prompting is the boundary where clearer wording alone cannot solve problems such as factuality, freshness, safety, or long-document consistency.
  • Why it matters: It separates prompt writing from system design. Some failures require retrieval, evaluation, approval, or human review rather than just better wording.
  • Related concepts: prompt, evaluation, hallucination
  • Core Section: P1-12.3

line plot

  • Meaning: A line plot connects values in sequence, often over time or iteration, so the reader can see how a quantity changes.
  • Why it matters: Line plots are the basic form for reading trends, training curves, function shapes, and changes over time.
  • Related concepts: plot, loss curve, axis
  • Core Section: P2-13.2
  • Appears in: P2-13.3

linear regression

  • Meaning: Linear regression models the relationship between input features and a continuous output with the simplest linear form first.
  • Why it matters: It gives a simple baseline for regression before more complex models are justified, and it supports coefficient-based interpretation.
  • Related concepts: regression, slope, residual
  • Core Section: P4-10.1
  • Appears in: P4-10.2, P4-10.3

linear structure

  • Meaning: A linear structure treats data as an ordered sequence where position and before-after relationships matter.
  • Why it matters: It distinguishes sequence-oriented structures such as arrays, lists, stacks, and queues from trees and graphs.
  • Related concepts: non-linear structure, array, data structure
  • Core Section: P2-9.1
  • Appears in: P2-9.4

list

  • Meaning: A list is an ordered collection of values that can be accessed by index and iterated over.
  • Why it matters: Lists are a basic programming container and a reference point for later distinctions between lists, arrays, vectors, and tensors.
  • Related concepts: index, dictionary, loop
  • Core Section: P2-8.2
  • Appears in: P2-8.3, P2-8.4, P2-8.7, P2-9.1

LLM

  • Meaning: An LLM is a large language model trained on large-scale text to predict and generate token sequences.
  • Why it matters: LLM is central to generative AI experience, but it should be separated from the full chatbot product, tool system, or all of AI.
  • Related concepts: generative AI, machine learning, deep learning, GPT, conversational LLM
  • Core Section: P1-1.3
  • Appears in: P1-9.3, P1-11.3, P6-19.1, P6-summary

loc

  • Meaning: loc is the Pandas selector used to choose rows and columns by labels rather than integer positions.
  • Why it matters: It helps readers distinguish selecting by semantic row or column labels from selecting by integer position with iloc.
  • Related concepts: index, column, iloc
  • Core Section: P2-12.2
  • Appears in: P2-12.3

local environment

  • Meaning: A local environment is the set of programs, files, packages, paths, and settings used to run code on the user's own computer.
  • Why it matters: The same code can behave differently across machines and settings. Local environment explains version, package, path, and virtual-environment issues.
  • Related concepts: runtime, virtual environment, working directory
  • Core Section: P2-7.1
  • Appears in: P2-7.4, P2-7.5, P2-10.2

log

  • Meaning: A log is a time-ordered record of what a system or run did, including inputs, steps, tool calls, responses, and errors.
  • Why it matters: Final answers rarely show why a run succeeded or failed. Logs support review, accountability, reproducibility, and operational debugging.
  • Related concepts: trace, harness, reproducibility, accountability
  • Core Section: P7-6.2
  • Appears in: P1-14.3, P1-14.5, P1-15.3, P7-6.1, P7-7.1, P7-7.2

logistic regression

  • Meaning: Logistic regression is a linear classification model that turns a linear score into a probability-like value for class decisions.
  • Why it matters: It is a simple, interpretable classification baseline and helps separate model score, probability-like output, threshold, and final class.
  • Related concepts: classification, threshold, decision boundary
  • Core Section: P4-11.1
  • Appears in: P4-11.2

long-context

  • Meaning: Long-context refers to the design problem of preserving and using information from a long input within one task.
  • Why it matters: Long inputs are not just a larger token limit. They require cost, latency, retrieval, summarization, and memory-management decisions.
  • Related concepts: context window, retrieval-augmented generation, RAG, sparse attention
  • Core Section: P6-4.3
  • Appears in: P5-14.2

long-term dependency

  • Meaning: Long-term dependency is the problem where information from much earlier in a sequence remains important but is hard for a sequential model to preserve.
  • Why it matters: It explains why basic recurrent models struggled with distant information and why LSTM, GRU, and attention became important.
  • Related concepts: RNN, recurrent neural network, LSTM, long short-term memory, Attention
  • Core Section: P5-12.2
  • Appears in: P1-9.3, P5-12.1, P5-13.1, P5-13.2, P5-14.1, P5-14.2

loop

  • Meaning: A loop repeats the same procedure over multiple values or steps without rewriting the same code each time.
  • Why it matters: Loops turn collections from stored data into repeated action and support later ideas such as preprocessing and training iteration.
  • Related concepts: iterable, iterator, list
  • Core Section: P2-8.4
  • Appears in: P2-8.5, P2-10.3

LoRA

  • Meaning: LoRA is an efficient adaptation method that keeps the base model mostly fixed and trains small additional update components.
  • Why it matters: It offers a practical middle ground between prompt-only adaptation and full fine-tuning, reducing training and storage cost.
  • Related concepts: fine-tuning, parameter, quantization
  • Core Section: P6-8.2
  • Appears in: P6-9.4, P6-9.5

loss curve

  • Meaning: A loss curve is a line plot showing how loss changes across training steps or epochs.
  • Why it matters: It reveals training flow, instability, plateauing, and overfitting more clearly than a single final loss value.
  • Related concepts: line plot, loss function, accuracy
  • Core Section: P2-13.2
  • Appears in: P2-13.3, P2-15.1

loss function

  • Meaning: A loss function turns prediction error into a number that training can try to reduce.
  • Why it matters: Training needs a numeric objective to reduce. The loss function defines what kind of error the model treats as costly.
  • Related concepts: error, objective function, gradient descent
  • Core Section: P2-6.2
  • Appears in: P2-6.3, P2-13.2, P2-15.1, P5-4.1, P5-4.2

LSTM, long short-term memory

  • Meaning: LSTM is a recurrent neural-network structure that uses gates to control what information to keep, forget, and update over a sequence.
  • Why it matters: It shows how recurrent models added memory-control mechanisms before attention became the dominant way to handle sequence context.
  • Related concepts: RNN, recurrent neural network, GRU, gated recurrent unit, hidden state
  • Core Section: P1-11.2
  • Appears in: P5-12.1, P5-12.2