© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Attention Out - 32k
    9. 3-ATT_32K-OAI
    11. 4150
    Prev
    Next
    INDEX
    Explanations

    phrases indicating initiation and influence, along with references to comparative sizes and conditions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-att_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-attn-out-SAEs/v5_32k_layer_3
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.3.hook_attn_out
    Hook Layer
    3
    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
    55
    +0.10
    0.4%
    108
    +0.09
    0.4%
    522
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    661
    +0.10
    0.04
    85
    +0.09
    0.03
    674
    +0.09
    0.04
    Negative Logits
    shoot
    -0.77
     DT
    -0.77
    raft
    -0.75
    lia
    -0.73
    gif
    -0.71
    lation
    -0.71
    landers
    -0.68
    DT
    -0.67
    imore
    -0.66
    units
    -0.65
    POSITIVE LOGITS
     prone
    0.71
     cloaked
    0.71
    ospons
    0.69
    ilant
    0.69
     wedd
    0.67
     disguised
    0.66
     outnumbered
    0.65
    ulnerable
    0.63
     incarn
    0.63
     advertis
    0.62
    Activations Density 0.314%

    No Known Activations