© 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. 10-RES_POST_32K-OAI
    11. 27184
    Prev
    Next
    INDEX
    Explanations

    variables and types related to programming, particularly in Rust

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    254
    +0.11
    0.5%
    420
    +0.10
    0.5%
    671
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    631
    +0.11
    0.03
    659
    +0.10
    0.04
    12
    +0.10
    0.03
    Negative Logits
    osate
    -0.73
     Mothers
    -0.71
    selves
    -0.65
     Consumers
    -0.65
    ivan
    -0.62
    Kings
    -0.61
    aeus
    -0.60
    hawks
    -0.59
     tablets
    -0.59
     Zamb
    -0.59
    POSITIVE LOGITS
    lem
    0.80
    ql
    0.77
    raction
    0.77
    igen
    0.77
    OF
    0.76
    uler
    0.73
    code
    0.70
    stack
    0.69
    ax
    0.69
    Arg
    0.67
    Activations Density 0.305%

    No Known Activations