© 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. 30617
    Prev
    Next
    INDEX
    Explanations

    underscores or placeholders in programming code

    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
    462
    +0.17
    0.7%
    302
    +0.10
    0.5%
    159
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    157
    +0.17
    0.04
    514
    +0.10
    0.04
    228
    +0.09
    0.04
    Negative Logits
    ONSORED
    -0.73
    iculture
    -0.73
    ioned
    -0.72
    ividual
    -0.70
    wagon
    -0.67
    inately
    -0.65
    Monitor
    -0.65
    zos
    -0.64
    webkit
    -0.64
     conduc
    -0.63
    POSITIVE LOGITS
     dash
    0.82
    attering
    0.75
     fate
    0.68
    oming
    0.67
     dime
    0.66
    endment
    0.66
     perspective
    0.66
    ©
    0.66
    fourth
    0.65
     cure
    0.65
    Activations Density 0.236%

    No Known Activations