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. 6463
    Prev
    Next
    INDEX
    Explanations

    actions related to publishing or distribution

    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
    270
    +0.19
    1.3%
    351
    +0.18
    1.2%
    706
    +0.17
    1.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    7
    +0.19
    0.01
    665
    +0.18
    0.01
    301
    +0.17
    0.00
    Negative Logits
    rax
    -0.83
    ffe
    -0.80
    etsk
    -0.80
    cius
    -0.76
    grave
    -0.75
     Aus
    -0.73
    ounty
    -0.73
    ortium
    -0.73
    ça
    -0.70
    agna
    -0.69
    POSITIVE LOGITS
    æĸ¹
    0.84
     widget
    0.80
    âĶģ
    0.77
     Reasons
    0.76
    ifice
    0.76
     HOUSE
    0.74
     Pengu
    0.74
     fetus
    0.74
     NETWORK
    0.74
    essee
    0.72
    Activations Density 0.003%

    No Known Activations