© 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. 8-RES_POST_32K-OAI
    11. 6240
    Prev
    Next
    INDEX
    Explanations

    the word "gan" and its variations, indicating a focus on the presence of this specific word

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.8.hook_resid_post
    Hook Layer
    8
    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
    +0.18
    0.8%
    528
    +0.13
    0.6%
    421
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.18
    0.03
    288
    +0.13
    0.03
    201
    +0.12
    0.02
    Negative Logits
    hovah
    -0.75
     Ambro
    -0.70
    atio
    -0.67
     Kore
    -0.66
     toile
    -0.65
     Canaver
    -0.65
    atile
    -0.65
    ributed
    -0.64
    onse
    -0.64
    von
    -0.63
    POSITIVE LOGITS
    uine
    0.92
    ãĥĥãĤ¯
    0.92
    ãģ¦
    0.84
    wich
    0.79
    20439
    0.79
    agher
    0.78
    cest
    0.77
    insula
    0.76
    forth
    0.76
    zyme
    0.75
    Activations Density 0.101%

    No Known Activations