© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 10-RES_POST_32K-OAI
    11. 29620
    Prev
    Next
    INDEX
    Explanations

    instances of the word "gran" and its variations, indicating a focus on size or granularity

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_10.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.10.hook_resid_post
    Hook Layer
    10
    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
    604
    +0.11
    0.5%
    156
    +0.10
    0.5%
    421
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    353
    +0.11
    0.01
    288
    +0.10
    0.01
    577
    +0.10
    0.01
    Negative Logits
     Lovecraft
    -0.72
    ORT
    -0.68
    yre
    -0.66
     Laugh
    -0.65
    chuk
    -0.64
     Conan
    -0.63
    ORK
    -0.63
    hered
    -0.61
    ean
    -0.61
    suits
    -0.61
    POSITIVE LOGITS
    ularity
    1.20
    ular
    1.14
    itic
    0.94
    ulated
    0.94
    ite
    0.91
    itary
    0.90
    ules
    0.89
    ulate
    0.88
    lund
    0.82
    vik
    0.81
    Activations Density 0.080%

    No Known Activations