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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlogSlackPrivacy & 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. 6452
    Prev
    Next
    INDEX
    Explanations

    references to election-related processes and activities

    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
    167
    +0.17
    1.0%
    251
    +0.15
    0.9%
    709
    +0.14
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    536
    +0.17
    0.01
    581
    +0.15
    0.01
    586
    +0.14
    0.01
    Negative Logits
     contag
    -0.75
     forg
    -0.70
     disappoint
    -0.69
     appetite
    -0.69
     bias
    -0.69
    wcs
    -0.69
     corrosion
    -0.68
     subtract
    -0.68
     distinguish
    -0.67
     punish
    -0.67
    POSITIVE LOGITS
    merce
    0.93
    enegger
    0.92
    issance
    0.91
     Cth
    0.90
    abouts
    0.90
    ossier
    0.89
    soDeliveryDate
    0.88
    emonium
    0.80
    WARD
    0.79
     helicop
    0.79
    Activations Density 0.009%

    No Known Activations