© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Residual Post - 32k
    9. 8-RES_POST_32K-OAI
    11. 7512
    Prev
    Next
    INDEX
    Explanations

    the word "Mot" and its variations, indicating a focus on motivation-related concepts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.8.hook_resid_post
    Hook Layer
    8
    Architecture
    standard
    Context Size
    64
    Dataset
    Skylion007/openwebtext
    Activation Function
    topk
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    314
    +0.14
    0.6%
    288
    +0.12
    0.6%
    347
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.14
    0.01
    585
    +0.12
    0.01
    746
    +0.11
    0.01
    Negative Logits
     Border
    -0.64
    theless
    -0.63
     Waters
    -0.62
     Gaul
    -0.60
    interstitial
    -0.60
     Silence
    -0.60
     Leopard
    -0.60
    holder
    -0.60
     WARN
    -0.60
     Rabbit
    -0.59
    POSITIVE LOGITS
    ivated
    1.35
    ivation
    1.32
    ivating
    1.22
    ör
    1.21
    oring
    1.11
    ierre
    1.06
    ility
    1.05
    oren
    1.05
    ypes
    1.04
    orm
    1.00
    Activations Density 0.046%

    No Known Activations