Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Attention Out - 128k
    5. 0-ATT_128K-OAI
    6. 13572
    Prev
    Next
    INDEX
    Explanations

    words related to succubi or themes of seduction and allure

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    jbloom/GPT2-Small-OAI-v5-128k-attn-out-SAEs/v5_128k_layer_0
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    131,072
    Data Type
    torch.float32
    Hook Name
    blocks.0.hook_attn_out
    Hook Layer
    0
    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
    151
    +0.19
    1.2%
    176
    +0.15
    1.0%
    566
    +0.15
    1.0%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    736
    +0.19
    0.00
    176
    +0.15
    0.01
    514
    +0.15
    0.01
    Negative Logits
    çīĪ
    -0.78
     GOODMAN
    -0.78
    abouts
    -0.77
     Strait
    -0.74
    raltar
    -0.74
    pocket
    -0.72
     Skydragon
    -0.72
     partName
    -0.71
    assetsadobe
    -0.69
    ibble
    -0.68
    POSITIVE LOGITS
    ³³³³
    0.83
    Joined
    0.79
    ³³³
    0.78
    ordes
    0.77
    Demon
    0.75
    initialized
    0.75
    à
    0.74
    URI
    0.73
    ");
    0.71
     Reincarnated
    0.70
    Activations Density 0.021%

    No Known Activations