© 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. 10-RES_POST_32K-OAI
    11. 29209
    Prev
    Next
    INDEX
    Explanations

    Quotation marks used in dialogues or citations

    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
    659
    +0.13
    0.6%
    749
    +0.13
    0.6%
    232
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.13
    0.03
    749
    +0.13
    0.02
    232
    +0.10
    0.02
    Negative Logits
    alysed
    -0.81
     apex
    -0.77
    artifacts
    -0.75
    */(
    -0.74
    inary
    -0.71
     facult
    -0.68
    ĻĤ
    -0.68
    worldly
    -0.68
    cffffcc
    -0.67
     coincide
    -0.66
    POSITIVE LOGITS
     Johnson
    1.12
     Andersen
    1.09
     Caldwell
    1.04
     Hart
    1.02
     Thompson
    1.01
     Shields
    0.99
     Welch
    0.97
     Jones
    0.97
     Bennett
    0.97
     Smith
    0.96
    Activations Density 0.104%

    No Known Activations