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. 6468
    Prev
    Next
    INDEX
    Explanations

    terms related to baseball and its associated terminology

    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
    508
    +0.21
    1.4%
    385
    +0.19
    1.3%
    167
    +0.19
    1.3%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    241
    +0.21
    0.02
    394
    +0.19
    0.03
    302
    +0.19
    0.02
    Negative Logits
    ippi
    -0.85
    etsk
    -0.84
    agnar
    -0.79
     Ukrain
    -0.75
    zech
    -0.73
    apego
    -0.71
     blo
    -0.71
    ukong
    -0.71
    imura
    -0.70
    imir
    -0.69
    POSITIVE LOGITS
     Tycoon
    0.87
     Clicker
    0.81
    reference
    0.74
    soDeliveryDate
    0.73
    IQ
    0.72
    SCP
    0.71
    Reader
    0.71
    CAST
    0.70
    iciency
    0.69
    lishes
    0.68
    Activations Density 0.124%

    No Known Activations