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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 13547
    Prev
    Next
    INDEX
    Explanations

    clauses related to terms and conditions

    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
    439
    +0.18
    1.3%
    575
    +0.17
    1.2%
    283
    +0.15
    1.1%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    52
    +0.18
    0.01
    575
    +0.17
    0.01
    642
    +0.15
    0.00
    Negative Logits
    iery
    -0.85
    aughters
    -0.83
    cyclopedia
    -0.79
    eph
    -0.77
    evin
    -0.75
    emy
    -0.73
     nomine
    -0.73
    inia
    -0.73
    icipated
    -0.73
    anmar
    -0.72
    POSITIVE LOGITS
     ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³
    1.07
    SPONSORED
    0.83
    ä
    0.79
     srfAttach
    0.78
     GEAR
    0.75
    é¾įå¥ij士
    0.71
    soDeliveryDate
    0.70
     Mechanical
    0.70
     dstg
    0.69
     Radiant
    0.69
    Activations Density 0.007%

    No Known Activations