© 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. 28402
    Prev
    Next
    INDEX
    Explanations

    punctuation marks and their context within sentences

    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
    373
    +0.11
    0.5%
    229
    +0.11
    0.5%
    75
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    6
    +0.11
    0.33
    640
    +0.11
    0.22
    145
    +0.10
    0.22
    Negative Logits
    ixty
    -0.74
    archives
    -0.72
     incent
    -0.72
    è¦ļéĨĴ
    -0.71
    pmwiki
    -0.69
    nown
    -0.67
    cedented
    -0.66
    eaturing
    -0.64
     TEXT
    -0.64
    userc
    -0.64
    POSITIVE LOGITS
     Hernandez
    1.26
     Rivera
    1.25
     Gallagher
    1.25
     Peterson
    1.23
     Kushner
    1.23
     Pryor
    1.23
     Robertson
    1.22
     McGregor
    1.22
     Keller
    1.21
     Olson
    1.20
    Activations Density 1.973%

    No Known Activations