© 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. 28473
    Prev
    Next
    INDEX
    Explanations

    instances of parentheses in the text

    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
    377
    +0.12
    0.5%
    496
    +0.10
    0.4%
    297
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    160
    +0.12
    0.13
    102
    +0.10
    0.09
    640
    +0.09
    0.07
    Negative Logits
     corrid
    -0.75
    ©¶æ
    -0.73
    ailability
    -0.69
     guiActiveUnfocused
    -0.66
     recept
    -0.65
     arri
    -0.65
     seiz
    -0.65
     terminated
    -0.64
     stabilized
    -0.63
    oldown
    -0.63
    POSITIVE LOGITS
    unless
    1.45
    especially
    1.41
    sorry
    1.39
    well
    1.37
    yet
    1.34
    maybe
    1.33
    yeah
    1.32
    although
    1.32
    hey
    1.30
    because
    1.29
    Activations Density 0.244%

    No Known Activations