© 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. 7107
    Prev
    Next
    INDEX
    Explanations

    punctuation marks, particularly commas

    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
    219
    +0.12
    0.6%
    616
    +0.10
    0.5%
    525
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    196
    +0.12
    0.03
    675
    +0.10
    0.03
    720
    +0.09
    0.03
    Negative Logits
     artif
    -0.66
    inventoryQuantity
    -0.64
     dime
    -0.64
     Andersen
    -0.60
    lication
    -0.58
     Perkins
    -0.57
     bearings
    -0.57
     DF
    -0.57
     gorilla
    -0.56
     Cav
    -0.56
    POSITIVE LOGITS
    leans
    0.81
    very
    0.81
    Ãį
    0.76
    ivably
    0.75
    few
    0.74
    ple
    0.66
    ivable
    0.65
     very
    0.65
    vere
    0.65
     VERY
    0.65
    Activations Density 0.225%

    No Known Activations