© 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 EvalsExportsinterp-engineNEWAPI 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. 8182
    Prev
    Next
    INDEX
    Explanations

    specific names or identifiers related to people or organizations

    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
    506
    +0.13
    0.6%
    664
    +0.10
    0.5%
    118
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.13
    0.05
    659
    +0.10
    0.04
    299
    +0.10
    0.04
    Negative Logits
     Calder
    -0.70
    rities
    -0.62
    rules
    -0.61
     Reboot
    -0.59
     endowed
    -0.59
     Visitors
    -0.59
     Bett
    -0.58
    erity
    -0.55
     Clover
    -0.55
     guiIcon
    -0.55
    POSITIVE LOGITS
    pta
    0.98
    ificant
    0.85
    atchewan
    0.84
    oslav
    0.82
    atoon
    0.76
    letal
    0.73
    ecast
    0.72
    pling
    0.72
    oku
    0.71
    olean
    0.70
    Activations Density 0.508%

    No Known Activations