© 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. 10-RES_POST_32K-OAI
    11. 28346
    Prev
    Next
    INDEX
    Explanations

    references to being alive or survival

    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
    132
    +0.13
    0.6%
    314
    +0.13
    0.6%
    223
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    759
    +0.13
    0.02
    468
    +0.13
    0.02
    762
    +0.12
    0.02
    Negative Logits
     Telescope
    -0.70
    RECT
    -0.67
    Americ
    -0.65
     Sorceress
    -0.63
    soDeliveryDate
    -0.63
    ãĥĩ
    -0.63
    PsyNetMessage
    -0.60
    argon
    -0.59
    ggles
    -0.59
     laun
    -0.58
    POSITIVE LOGITS
    lihood
    0.92
    lier
    0.85
    beat
    0.84
    abouts
    0.80
     enough
    0.78
    spin
    0.77
    guard
    0.75
    zai
    0.74
    weight
    0.73
    guards
    0.72
    Activations Density 0.094%

    No Known Activations