© 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. 28562
    Prev
    Next
    INDEX
    Explanations

    variations of the word "ir"

    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
    238
    +0.16
    0.7%
    314
    +0.10
    0.5%
    745
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    238
    +0.16
    0.03
    284
    +0.10
    0.02
    314
    +0.09
    -0.00
    Negative Logits
    e
    -0.66
     untreated
    -0.60
     compliment
    -0.60
    er
    -0.59
     solvent
    -0.59
    eur
    -0.58
     dial
    -0.58
     socket
    -0.57
     Weber
    -0.55
     duct
    -0.55
    POSITIVE LOGITS
    vana
    1.16
    acial
    1.04
    mingham
    1.02
    rh
    0.96
    apt
    0.92
    abbit
    0.92
    rha
    0.90
    ror
    0.90
    nel
    0.89
    iously
    0.89
    Activations Density 0.154%

    No Known Activations