© 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. 6202
    Prev
    Next
    INDEX
    Explanations

    references to academic papers and scholarly citations

    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
    373
    +0.23
    1.1%
    576
    +0.10
    0.5%
    488
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.23
    0.11
    232
    +0.10
    0.08
    373
    +0.09
    -0.04
    Negative Logits
     alike
    -0.91
    tops
    -0.60
    terday
    -0.58
     rave
    -0.56
    whatever
    -0.56
     traverse
    -0.54
     smack
    -0.54
     whis
    -0.53
     Scotch
    -0.53
     kick
    -0.52
    POSITIVE LOGITS
    ibliography
    0.88
     âĵĺ
    0.76
    ":["
    0.76
    tymology
    0.70
    ptin
    0.69
    malink
    0.67
    ukong
    0.67
     largeDownload
    0.66
     GOODMAN
    0.66
     CrossRef
    0.65
    Activations Density 0.771%

    No Known Activations