© 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. 6731
    Prev
    Next
    INDEX
    Explanations

    instances of the word "while."

    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
    476
    +0.11
    0.5%
    738
    +0.11
    0.5%
    294
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    294
    +0.11
    0.04
    172
    +0.11
    0.04
    476
    +0.10
    0.04
    Negative Logits
    ahime
    -0.80
    oux
    -0.73
    aer
    -0.71
    eri
    -0.69
    ISE
    -0.69
    imentary
    -0.67
    enture
    -0.66
    etic
    -0.65
    iola
    -0.65
    backer
    -0.64
    POSITIVE LOGITS
     awaiting
    0.91
     browsing
    0.91
     traveling
    0.89
     maintaining
    0.89
     retaining
    0.88
     others
    0.88
     watching
    0.86
     researching
    0.85
     simultaneously
    0.84
     preserving
    0.83
    Activations Density 0.128%

    No Known Activations