© 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
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 8-RES_POST_32K-OAI
    6. 4530
    Prev
    Next
    INDEX
    Explanations

    references to the word "with."

    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
    362
    +0.21
    1.0%
    393
    +0.18
    0.9%
    608
    +0.15
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.21
    0.13
    60
    +0.18
    0.11
    608
    +0.15
    0.12
    Negative Logits
    hood
    -0.75
    sburg
    -0.73
     with
    -0.72
    NN
    -0.72
    ghai
    -0.70
    sylvania
    -0.69
    hops
    -0.64
    pad
    -0.64
    fair
    -0.64
    sat
    -0.64
    POSITIVE LOGITS
     regard
    1.70
     regards
    1.69
    stood
    1.57
     respect
    1.45
    draw
    1.45
     impunity
    1.40
    drawn
    1.26
    standing
    1.22
    holding
    1.10
     utmost
    1.01
    Activations Density 1.104%

    No Known Activations