© 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. 7045
    Prev
    Next
    INDEX
    Explanations

    the occurrences of the name "She" in various contexts

    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
    218
    +0.13
    0.6%
    174
    +0.12
    0.6%
    610
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    610
    +0.13
    0.02
    235
    +0.12
    0.03
    612
    +0.09
    0.03
    Negative Logits
     conditioning
    -0.64
     reinforcement
    -0.63
     prelim
    -0.61
     viz
    -0.61
    é¾įå¥ij士
    -0.61
    ãĥ³ãĤ¸
    -0.61
     impressions
    -0.60
    ument
    -0.59
     inst
    -0.57
     citation
    -0.57
    POSITIVE LOGITS
    pherd
    1.37
    ffield
    1.15
    pard
    1.11
    athing
    1.08
    lf
    1.06
    ldon
    1.03
    arer
    0.99
    aring
    0.98
    eper
    0.98
    lling
    0.96
    Activations Density 0.065%

    No Known Activations