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

    Neuronpedia

    Jacobian LensNEW
    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. Residual Post - 32k
    9. 10-RES_POST_32K-OAI
    11. 29045
    Prev
    Next
    INDEX
    Explanations

    expressions of affirmation or emphasis

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    203
    +0.13
    0.6%
    219
    +0.09
    0.4%
    449
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    240
    +0.13
    0.06
    715
    +0.09
    0.06
    155
    +0.09
    0.05
    Negative Logits
     ®
    -0.63
     awaits
    -0.60
    Optional
    -0.60
     Bung
    -0.58
    quart
    -0.57
     threatens
    -0.57
    SourceFile
    -0.56
     adherent
    -0.55
     vanished
    -0.54
     Moderate
    -0.53
    POSITIVE LOGITS
    entimes
    0.95
     thank
    0.85
     anytime
    0.76
     whenever
    0.74
    FTWARE
    0.70
    when
    0.70
     yeah
    0.69
     many
    0.68
    sometimes
    0.68
    thank
    0.68
    Activations Density 0.170%

    No Known Activations