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. 6471
    Prev
    Next
    INDEX
    Explanations

    patterns and symbols related to technical or programming contexts

    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
    417
    +0.20
    1.3%
    270
    +0.17
    1.1%
    706
    +0.16
    1.1%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    160
    +0.20
    0.08
    124
    +0.17
    0.06
    4
    +0.16
    0.06
    Negative Logits
    ãĤ©
    -0.82
     Vessel
    -0.72
     MPG
    -0.71
     Solo
    -0.69
     Seller
    -0.67
    ogether
    -0.66
     Haz
    -0.65
     Kidd
    -0.64
     suppose
    -0.62
     mine
    -0.61
    POSITIVE LOGITS
    Accessory
    0.90
    domain
    0.88
    ItemTracker
    0.84
    task
    0.77
    handled
    0.76
    ************
    0.74
    SIGN
    0.74
    qa
    0.73
    atomic
    0.70
    communications
    0.69
    Activations Density 0.356%

    No Known Activations