Search

www.springer.com The European Mathematical Society

  • StatProb Collection
  • Recent changes
  • Current events
  • Random page
  • Project talk
  • Request account
  • What links here
  • Related changes
  • Special pages
  • Printable version
  • Permanent link
  • Page information
  • View source

Linear representation

A homomorphism $\pi$ of a group (respectively an algebra, ring, semi-group) $X$ into the group of all invertible linear operators on a vector space $E$ (respectively, into the algebra, ring, multiplicative semi-group of all linear operators on $E$). If $E$ is a topological vector space, then a linear representation of $X$ on $E$ is a representation whose image contains only continuous linear operators on $E$. The space $E$ is called the representation space of $\pi$ and the operators $\pi(x)$, $x\in X$, are called the operators of the representation $\pi$.

  • Group theory and generalizations
  • This page was last edited on 25 October 2014, at 19:18.
  • Privacy policy
  • About Encyclopedia of Mathematics
  • Disclaimers
  • Impressum-Legal

Help | Advanced Search

Computer Science > Computation and Language

Title: the linear representation hypothesis and the geometry of large language models.

Abstract: Informally, the 'linear representation hypothesis' is the idea that high-level concepts are represented linearly as directions in some representation space. In this paper, we address two closely related questions: What does "linear representation" actually mean? And, how do we make sense of geometric notions (e.g., cosine similarity or projection) in the representation space? To answer these, we use the language of counterfactuals to give two formalizations of "linear representation", one in the output (word) representation space, and one in the input (sentence) space. We then prove these connect to linear probing and model steering, respectively. To make sense of geometric notions, we use the formalization to identify a particular (non-Euclidean) inner product that respects language structure in a sense we make precise. Using this causal inner product, we show how to unify all notions of linear representation. In particular, this allows the construction of probes and steering vectors using counterfactual pairs. Experiments with LLaMA-2 demonstrate the existence of linear representations of concepts, the connection to interpretation and control, and the fundamental role of the choice of inner product.

Submission history

Access paper:.

  • Other Formats

license icon

References & Citations

  • Google Scholar
  • Semantic Scholar

BibTeX formatted citation

BibSonomy logo

Bibliographic and Citation Tools

Code, data and media associated with this article, recommenders and search tools.

  • Institution

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs .

nLab linear representation

Linear algebra.

homotopy theory , (∞,1)-category theory , homotopy type theory

flavors: stable , equivariant , rational , p-adic , proper , geometric , cohesive , directed …

models: topological , simplicial , localic , …

see also algebraic topology

Introductions

Introduction to Basic Homotopy Theory

Introduction to Abstract Homotopy Theory

geometry of physics – homotopy types

Definitions

homotopy , higher homotopy

homotopy type

Pi-algebra , spherical object and Pi(A)-algebra

homotopy coherent category theory

homotopical category

model category

category of fibrant objects , cofibration category

Waldhausen category

homotopy category

(∞,1)-category

  • homotopy category of an (∞,1)-category

Paths and cylinders

left homotopy

cylinder object

mapping cone

right homotopy

path object

mapping cocone

universal bundle

interval object

homotopy localization

infinitesimal interval object

Homotopy groups

homotopy group

fundamental group

  • fundamental group of a topos

Brown-Grossman homotopy group

categorical homotopy groups in an (∞,1)-topos

geometric homotopy groups in an (∞,1)-topos

fundamental ∞-groupoid

fundamental groupoid

  • path groupoid

fundamental ∞-groupoid in a locally ∞-connected (∞,1)-topos

fundamental ∞-groupoid of a locally ∞-connected (∞,1)-topos

fundamental (∞,1)-category

  • fundamental category

Basic facts

  • fundamental group of the circle is the integers

fundamental theorem of covering spaces

Freudenthal suspension theorem

Blakers-Massey theorem

higher homotopy van Kampen theorem

nerve theorem

Whitehead's theorem

Hurewicz theorem

Galois theory

homotopy hypothesis -theorem

Characters of linear representations

Characteristic classes of linear representations, related concepts.

A linear representation is a representation on a category of vector spaces or similar ( Vect , Mod , etc.)

This is the most common flavor of representations . One sometimes considers representations on objects other than linear spaces (such as permutation representations ) but often these are called not representations but actions .

See characters of linear representations .

Under the Atiyah-Segal completion map linear representations of a group G G induce topological K-theory classes on the classifying space B G B G . Their Chern classes are hence invariants of the linear representations themselves.

See at characteristic class of a linear representation for more.

permutation representation

category of representations

  • Tammo tom Dieck , Representation theory , 2009 ( pdf )

For more see the references at representation theory .

Last revised on February 1, 2019 at 14:18:03. See the history of this page for a list of all contributions to it.

Linear Representation: Contemporary Use

  • Conference paper
  • First Online: 18 March 2020
  • Cite this conference paper

Book cover

  • Tommaso Empler 15  

Part of the book series: Advances in Intelligent Systems and Computing ((AISC,volume 1140))

Included in the following conference series:

  • International and Interdisciplinary Conference on Image and Imagination

1350 Accesses

“Linear representation” has been used, in the past of western history and near east, with celebrative functions and a mode of representation where a perception of depth is used linked to the overlapping of elements present in the scene and a consequent visualization in side elevation. A rediscovery of “linear representation” occurs in contemporary age, where it assumes a technical use in temporary and permanent environmental solutions or in new media, defining a new research field for representation applied to architecture, graphics and design, better known as “environmental graphics”. In this context, a particular development takes place on some ephemeral structures such as yard fences, with horizontal development, or in facades of buildings.

  • Linear representation
  • Environmental graphic
  • Perceptive depth cues
  • Gestalt principles

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
  • Available as EPUB and PDF
  • Compact, lightweight edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Costantin, A.L.: Compendio di geometria pratica applicata al disegno lineare. Officina tipografica di Giovanni Enrici e comp., Saluzzo (1843)

Google Scholar  

De Fiore, G.: Dizionario del Disegno. La Scuola Editore, Brescia (1967)

Empler, T.: Grafica e comunicazione ambientale. Dei – Tipografia del genio Civile, Roma (2012)

Gagnazzo, S.:. Muro di Berlino, tutta la sua street art (2017). https://www.lastampa.it/2017/07/28/cultura/muro-di-berlino-tutta-la-sua-street-art-9EG0CAfluiHUUHoDlVY6EP/pagina.html . Accessed 12 Feb 2019

Norberg-Schulz, C., Norberg-Schulz, A.M.: Genius Loci. Paesaggio, ambiente, architettura. Electa - Documenti di architettura, Milano (1992)

Schreiber, G.: Il disegno lineare corso pratico per artisti e industriali e specialmente per le scuole tecniche normali e professionali. Loescher, Torino (1872)

Valeri, V.: Corso di Disegno 2. La Nuova Italia Editrice, Firenze (1990)

Vallortigara, G., Panciera, N.: Cervelli che contano. Adelphi Edizioni, Milano (2014)

Download references

Author information

Authors and affiliations.

Department of History, Representation and Restoration of Architecture, Sapienza University of Rome, Piazza Borghese 9, 00186, Rome, Italy

Tommaso Empler

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Tommaso Empler .

Editor information

Editors and affiliations.

University of Sassari, Alghero, Italy

Enrico Cicalò

Rights and permissions

Reprints and permissions

Copyright information

© 2020 Springer Nature Switzerland AG

About this paper

Cite this paper.

Empler, T. (2020). Linear Representation: Contemporary Use. In: Cicalò, E. (eds) Proceedings of the 2nd International and Interdisciplinary Conference on Image and Imagination. IMG 2019. Advances in Intelligent Systems and Computing, vol 1140. Springer, Cham. https://doi.org/10.1007/978-3-030-41018-6_40

Download citation

DOI : https://doi.org/10.1007/978-3-030-41018-6_40

Published : 18 March 2020

Publisher Name : Springer, Cham

Print ISBN : 978-3-030-41017-9

Online ISBN : 978-3-030-41018-6

eBook Packages : Engineering Engineering (R0)

Share this paper

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

  • Publish with us

Policies and ethics

  • Find a journal
  • Track your research

LISR: Learning Linear 3D Implicit Surface Representation Using Compactly Supported Radial Basis Functions

  • Atharva Pandey Indian Institute of Technology, Jodhpur
  • Vishal Yadav Indian Institute of Technology, Jodhpur
  • Rajendra Nagar Indian Institute of Technology, Jodhpur
  • Santanu Chaudhury Indian Institute of Technology, Jodhpur

AAAI-24 / IAAI-24 / EAAI-24 Proceedings Cover

  • Video/Poster

How to Cite

  • Endnote/Zotero/Mendeley (RIS)

Information

  • For Readers
  • For Authors
  • For Librarians

Developed By

Subscription.

Login to access subscriber-only resources.

Part of the PKP Publishing Services Network

Copyright © 2024, Association for the Advancement of Artificial Intelligence

More information about the publishing system, Platform and Workflow by OJS/PKP.

Library homepage

  • school Campus Bookshelves
  • menu_book Bookshelves
  • perm_media Learning Objects
  • login Login
  • how_to_reg Request Instructor Account
  • hub Instructor Commons
  • Download Page (PDF)
  • Download Full Book (PDF)
  • Periodic Table
  • Physics Constants
  • Scientific Calculator
  • Reference & Cite
  • Tools expand_more
  • Readability

selected template will load here

This action is not available.

Mathematics LibreTexts

3.2.1: Introduction to Linear Relationships

  • Last updated
  • Save as PDF
  • Page ID 35945

  • Illustrative Mathematics
  • OpenUp Resources

Let's explore some relationships between two variables.

Exercise \(\PageIndex{1}\): Number Talk: Fraction Division

Find the value of \(2\frac{5}{8}\div\frac{1}{2}\).

Exercise \(\PageIndex{2}\): Stacking Cups

We have two stacks of Styrofoam cups.

  • One stack has 6 cups, and its height is 15 cm.
  • The other stack has 12 cups, and its height is 23 cm.

How many cups are needed for a stack with a height of 50 cm?

clipboard_ea9703c80c79d88b9a5ac472bb9c38a04.png

Exercise \(\PageIndex{3}\): Connecting Slope to Rate of Change

1. If you didn’t create your own graph of the situation before, do so now.

clipboard_e80422c6af936c2f6bbac2d05580d3281.png

2. What are some ways you can tell that the number of cups is not proportional to the height of the stack?

3. What is the slope of the line in your graph? What does the slope mean in this situation?

4. At what point does your line intersect the vertical axis? What do the coordinates of this point tell you about the cups?

5. How much height does each cup after the first add to the stack?

Andre starts babysitting and charges $10 for traveling to and from the job, and $15 per hour. For every additional hour he works he charges another $15. If we graph Andre's earnings based on how long he works, we have a line that starts at $10 on the vertical axis and then increases by $15 each hour. A linear relationship is any relationship between two quantities where one quantity has a constant rate of change with respect to the other.

clipboard_e66f3e53b0df44187243ec500cea32566.png

We can figure out the rate of change using the graph. Because the rate of change is constant, we can take any two points on the graph and divide the amount of vertical change by the amount of horizontal change. For example, take the points \((2,40)\) and \((6,100)\). They mean that Andre earns $40 for working 2 hours and $100 for working 6 hours. The rate of change is \(\frac{100-40}{6-2}=15\) dollars per hour. Andre's earnings go up $15 for each hour of babysitting.

Notice that this is the same way we calculate the slope of the line. That's why the graph is a line, and why we call this a linear relationship. The rate of change of a linear relationship is the same as the slope of its graph.

With proportional relationships we are used to graphs that contain the point \((0,0)\). But proportional relationships are just one type of linear relationship. In the following lessons, we will continue to explore the other type of linear relationship where the quantities are not both 0 at the same time.

Glossary Entries

Definition: Linear Relationship

A linear relationship between two quantities means they are related like this: When one quantity changes by a certain amount, the other quantity always changes by a set amount. In a linear relationship, one quantity has a constant rate of change with respect to the other.

The relationship is called linear because its graph is a line.

The graph shows a relationship between number of days and number of pages read.

When the number of days increases by 2, the number of pages read always increases by 60. The rate of change is constant, 30 pages per day, so the relationship is linear.

clipboard_e4d1fd0f0259d7198073a02ce75a7303d.png

Exercise \(\PageIndex{4}\)

A restaurant offers delivery for their pizzas. The total cost is a delivery fee added to the price of the pizzas. One customer pays $25 to have 2 pizzas delivered. Another customer pays $58 for 5 pizzas. How many pizzas are delivered to a customer who pays $80?

Exercise \(\PageIndex{5}\)

To paint a house, a painting company charges a flat rate of $500 for supplies, plus $50 for each hour of labor.

  • How much would the painting company charge to paint a house that needs 20 hours of labor? A house that needs 50 hours?
  • Draw a line representing the relationship between \(x\), the number of hours it takes the painting company to finish the house, and \(y\), the total cost of painting the house. Label the two points from the earlier question on your graph.

clipboard_e2f3f04e7728753811d13dd41e53c57de.png

3. Find the slope of the line. What is the meaning of the slope in this context?

Exercise \(\PageIndex{6}\)

Tyler and Elena are on the cross country team.

Tyler's distances and times for a training run are shown on the graph.

clipboard_ef7eeb5c0bf090d609200590b4107f2fa.png

Elena’s distances and times for a training run are given by the equation \(y=8.5x\), where \(x\) represents distance in miles and \(y\) represents time in minutes.

  • Who ran farther in 10 minutes? How much farther? Explain how you know.
  • Calculate each runner's pace in minutes per mile.
  • Who ran faster during the training run? Explain or show your reasoning.

(From Unit 3.1.4)

Exercise \(\PageIndex{7}\)

Write an equation for the line that passes through \((2,5)\) and \((6,7)\).

(From Unit 2.3.3)

COMMENTS

  1. PDF Introduction to Representation Theory

    Definition 1.2.Suppose Gis a group. A linear representation of Gis a group homomorphism ρ: G→GL(V) for a vector space V. The degree of a linear representation is the dimension of the corresponding vector space V. A linear representation ρis faithful if ρis injective as a map of sets.

  2. Group representation

    A representation of a group "acts" on an object. A simple example is how the symmetries of a regular polygon, consisting of reflections and rotations, transform the polygon.. In the mathematical field of representation theory, group representations describe abstract groups in terms of bijective linear transformations of a vector space to itself (i.e. vector space automorphisms); in particular ...

  3. PDF Representation Theory

    meaning that we have an equality of linear maps after inserting any group element g. Observe that this relation determines ρ 2, if ρ 1 and φare known. We can finally formulate the Basic Problem of Representation Theory: Classify all representations of a given group G, up to isomorphism. For arbitrary G, this is very hard!

  4. PDF LINEAR REPRESENTATIONS OF FINITE GROUPS

    Definition 2.5.Let V be a vector space and W be a subspace of V. If ρ: G→ GL(V) is a linear representation, and if W is invariant under the action of G, or ρ(g)w∈W for all g∈G, w∈W, then the restriction of ρto W, ρW, is a linear representation. W is then called a subrepresentation of V. There are always two

  5. Linear representation

    [a1] C.W. Curtis, I. Reiner, "Methods of representation theory" , 1-2, Wiley (Interscience) (1981-1987) [a2] J.-P. Serre, "Répresentations linéaires des groupes ...

  6. PDF Introduction to representation theory

    Very roughly speaking, representation theory studies symmetry in linear spaces. It is a beautiful mathematical subject which has many applications, ranging from number theory and combinatorics to ge-ometry, probability theory, quantum mechanics, and quantum eld theory. Representation theory was born in 1896 in the work of the Ger-

  7. 2.1: Vectors and Linear Combinations

    Proposition 2.1.7. The vector b is a linear combination of the vectors v1, v2, …, vn if and only if the linear system corresponding to the augmented matrix. [v1 v2 … vn b] is consistent. A solution to this linear system gives weights c1, c2, …, cn such that. c1v1 + c2v2 + … + cnvn = b.

  8. 3.2.3: Representations of Linear Relationships

    The relationship is called linear because its graph is a line. The graph shows a relationship between number of days and number of pages read. When the number of days increases by 2, the number of pages read always increases by 60. The rate of change is constant, 30 pages per day, so the relationship is linear. Figure 3.2.3.5 3.2.3. 5.

  9. The Linear Representation Hypothesis and the ...

    This paper addresses two closely related questions: What does "linear representation"actually mean? Informally, the 'linear representation hypothesis' is the idea that high-level concepts are represented linearly as directions in some representation space. In this paper, we address two closely related questions: What does"linear representation"actually mean? And, how do we make sense of ...

  10. PDF Generalities on linear representations

    linear representations Let V be a vector space over the field C of complex numbers and let GL(V) be the group of isomorphisms of V onto itself. An element a of GL(V) is, by definition, a linear mapping of V into V which has an inverse a-I; this inverse is linear. When V has a finite basis (ej) of n elements, each linear

  11. Interpreting linear models

    What is a linear model? If we graph data and notice a trend that is approximately linear, we can model the data with a line of best fit. A line of best fit can be estimated by drawing a line so that the number of points above and below the line is about equal. 1 2 3 2 4 6 8 10 y x 0 0. While a line of best fit is not an exact representation of ...

  12. The Linear Representation Hypothesis and the Geometry of ...

    Informally, the 'linear representation hypothesis' is the idea that high-level concepts are represented linearly as directions in some representation space. In this paper, we address two closely related questions: What does "linear representation" actually mean? And, how do we make sense of geometric notions (e.g., cosine similarity

  13. Representation theory

    Representation theory studies how algebraic structures "act" on objects. A simple example is how the symmetries of regular polygons, consisting of reflections and rotations, transform the polygon.. Representation theory is a branch of mathematics that studies abstract algebraic structures by representing their elements as linear transformations of vector spaces, and studies modules over these ...

  14. [2311.03658] The Linear Representation Hypothesis and the Geometry of

    The Linear Representation Hypothesis and the Geometry of Large Language Models. Informally, the 'linear representation hypothesis' is the idea that high-level concepts are represented linearly as directions in some representation space. In this paper, we address two closely related questions: What does "linear representation" actually mean?

  15. linear representation in nLab

    A linear representation is a representation on a category of vector spaces or similar ( Vect, Mod, etc.) This is the most common flavor of representations. One sometimes considers representations on objects other than linear spaces (such as permutation representations) but often these are called not representations but actions.

  16. Linear Representation

    1. Linear representations We will refer to representations of data based on symbols (language or knowledge representation) as linear representations. In situations in which such representations are available for the domain of the desired image, they may be processed to obtain information on how to choose the artifacts.

  17. Understanding linear relationships

    A linear relationship is any relationship between two variables that creates a line when graphed in the x y -plane. Linear relationships are very common in everyday life. Example: Maya and Geoff's heights. Maya is 3 inches taller than Geoff. linear relationship y x. Maya's height = Geoff's height + 3 y = x + 3. Example: Tai's runs.

  18. Linear Representation: Contemporary Use

    Linear representation is used, however, directly in the field of mathematics and physics, where the meaning is associated with the variability in measure and proportion of physical quantity taken into consideration (Vallortigara and Panciera 2014), or in cognitive logic or psychology, with reference to mental maps, seen, in some cases, as a ...

  19. 4.1: Linear Functions

    Definition: Linear Function. A linear function is a function whose graph is a line. Linear functions can be written in the slope-intercept form of a line. f(x) = mx + b. where b is the initial or starting value of the function (when input, x = 0 ), and m is the constant rate of change, or slope of the function. The y-intercept is at (0, b).

  20. Linear equations, functions, & graphs

    Test your understanding of Linear equations, functions, & graphs with these NaN questions. Start test. This topic covers: - Intercepts of linear equations/functions - Slope of linear equations/functions - Slope-intercept, point-slope, & standard forms - Graphing linear equations/functions - Writing linear equations/functions - Interpreting ...

  21. 2.1: Linear Functions

    Definition: Linear Function. A linear function is a function whose graph is a line. Linear functions can be written in the slope-intercept form of a line. \ [f (x)=mx+b\] where \ (b\) is the initial or starting value of the function (when input, \ (x=0\)), and \ (m\) is the constant rate of change, or slope of the function.

  22. Meaning of Riesz representations in a layman's term?

    This follows from the definition of the inner product and that the Hilbert space is complete. It does not require the Riesz representation theorem. It is the other direction -- from functional to an element in the Hilbert space that requires the functional to be bounded and linear in the general case. $\endgroup$ -

  23. LISR: Learning Linear 3D Implicit Surface Representation Using

    Implicit 3D surface reconstruction of an object from its partial and noisy 3D point cloud scan is the classical geometry processing and 3D computer vision problem. In the literature, various 3D shape representations have been developed, differing in memory efficiency and shape retrieval effectiveness, such as volumetric, parametric, and implicit surfaces.

  24. 3.2.1: Introduction to Linear Relationships

    In a linear relationship, one quantity has a constant rate of change with respect to the other. The relationship is called linear because its graph is a line. The graph shows a relationship between number of days and number of pages read. When the number of days increases by 2, the number of pages read always increases by 60.