© 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. 10-RES_POST_32K-OAI
    11. 28166
    Prev
    Next
    INDEX
    Explanations

    references to the term "globe" or its variations

    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
    522
    +0.11
    0.5%
    553
    +0.11
    0.5%
    459
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.11
    0.03
    526
    +0.11
    0.01
    659
    +0.11
    0.02
    Negative Logits
    taining
    -0.65
    ÄŁ
    -0.64
    TAIN
    -0.61
    FINEST
    -0.60
     Debor
    -0.60
    hyde
    -0.60
    cles
    -0.58
    ched
    -0.58
    riage
    -0.57
     Intelligent
    -0.57
    POSITIVE LOGITS
    acial
    1.00
    obl
    0.97
    ength
    0.95
    ossier
    0.95
    adiator
    0.94
    ossip
    0.94
    ibrary
    0.90
    ocal
    0.90
    uten
    0.90
    ocate
    0.89
    Activations Density 0.214%

    No Known Activations