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. 4218
    Prev
    Next
    INDEX
    Explanations

    instances of the word "away."

    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
    420
    +0.14
    0.7%
    410
    +0.12
    0.5%
    136
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    501
    +0.14
    0.02
    174
    +0.12
    0.02
    666
    +0.10
    0.02
    Negative Logits
    regor
    -0.77
    manship
    -0.70
    tyard
    -0.68
    irection
    -0.67
    wark
    -0.64
    ļéĨĴ
    -0.63
    rador
    -0.62
    ARDIS
    -0.62
    iors
    -0.61
    ruff
    -0.61
    POSITIVE LOGITS
     reckoning
    0.75
     peacefully
    0.75
     unsc
    0.71
     happily
    0.64
     clutching
    0.64
     sparing
    0.63
    happy
    0.63
     miser
    0.63
    ennes
    0.63
    ocaust
    0.62
    Activations Density 0.078%

    No Known Activations