© 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 EvalsExportsAPI 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. 6432
    Prev
    Next
    INDEX
    Explanations

    instances of direct speech or quotations

    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
    107
    +0.17
    0.8%
    669
    +0.13
    0.6%
    257
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    107
    +0.17
    0.07
    6
    +0.13
    0.06
    246
    +0.11
    0.05
    Negative Logits
    senal
    -0.97
     indemn
    -0.74
     synthes
    -0.70
     visual
    -0.69
     microbiome
    -0.69
     shading
    -0.68
     conceptual
    -0.68
     ont
    -0.67
     reactive
    -0.67
     particle
    -0.66
    POSITIVE LOGITS
    But
    1.29
    Otherwise
    1.29
    Plus
    1.27
    And
    1.22
    That
    1.17
    They
    1.17
    Then
    1.14
    Especially
    1.14
    Therefore
    1.14
    It
    1.12
    Activations Density 0.126%

    No Known Activations