© 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. 6664
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "in" as well as references to historical events and discussions

    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
    349
    +0.12
    0.5%
    2
    +0.09
    0.4%
    713
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.12
    0.22
    608
    +0.09
    0.18
    22
    +0.09
    0.11
    Negative Logits
    */(
    -0.86
    =-=-
    -0.73
    $$$$
    -0.67
    POSE
    -0.66
    LED
    -0.66
    ospace
    -0.66
    Widget
    -0.64
    asive
    -0.63
    ILA
    -0.63
    Ingredients
    -0.62
    POSITIVE LOGITS
     1980
    1.46
     1990
    1.42
     1970
    1.40
     1989
    1.39
     1960
    1.37
     1968
    1.37
     1975
    1.36
     1969
    1.35
     1986
    1.35
     1996
    1.34
    Activations Density 1.889%

    No Known Activations