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. 6433
    Prev
    Next
    INDEX
    Explanations

    specific names or terms related to "Bre" paired with significant occurrences or events

    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
    526
    +0.27
    2.1%
    733
    +0.18
    1.4%
    377
    +0.16
    1.2%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    624
    +0.27
    0.01
    57
    +0.18
    0.01
    326
    +0.16
    0.01
    Negative Logits
    ãĥį
    -0.93
    6666
    -0.85
    Reply
    -0.83
    cheon
    -0.82
     //[
    -0.77
     Bulldogs
    -0.75
    ãĥŁ
    -0.75
     Forestry
    -0.75
    Operation
    -0.72
    Pakistan
    -0.71
    POSITIVE LOGITS
    htaking
    1.10
    ngth
    1.10
    icter
    1.06
     livest
    0.95
    eport
    0.91
     psychiat
    0.88
     princ
    0.87
     challeng
    0.85
    uters
    0.84
     undermin
    0.84
    Activations Density 0.011%

    No Known Activations