Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 6-RES_POST_32K-OAI
    6. 4216
    Prev
    Next
    INDEX
    Explanations

    the names "Bob" and "Hay" in varying frequencies

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 6-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_6.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.6.hook_resid_post
    Hook Layer
    6
    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
    679
    +0.16
    0.7%
    699
    +0.12
    0.6%
    286
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.16
    0.03
    232
    +0.12
    0.02
    286
    +0.12
    0.02
    Negative Logits
    ãĥ¥
    -0.91
    istically
    -0.82
    umbers
    -0.81
    cer
    -0.80
    antly
    -0.78
    ific
    -0.78
    cers
    -0.78
    llular
    -0.76
    uously
    -0.76
    ¯¯¯¯
    -0.75
    POSITIVE LOGITS
    ¿½
    0.90
     ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³
    0.80
    sonian
    0.77
    anke
    0.75
    tie
    0.73
    NESS
    0.69
    maker
    0.67
    ede
    0.67
    riad
    0.66
    ãĤ¡
    0.65
    Activations Density 0.071%

    No Known Activations