© 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 EvalsExportsinterp-engineNEWAPI 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. 27864
    Prev
    Next
    INDEX
    Explanations

    occurrences of the article "a" and other related articles

    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
    756
    +0.16
    0.7%
    314
    +0.12
    0.5%
    648
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    36
    +0.16
    0.09
    250
    +0.12
    0.10
    328
    +0.10
    0.07
    Negative Logits
     unlucky
    -0.86
     Qiao
    -0.78
     awfully
    -0.75
     awful
    -0.74
     bunch
    -0.73
     tricky
    -0.73
     eerie
    -0.73
     nice
    -0.73
     wee
    -0.71
    prisingly
    -0.71
    POSITIVE LOGITS
     duly
    0.89
     lawfully
    0.88
     dwelling
    0.86
     bona
    0.86
     licensee
    0.86
     lawful
    0.85
     Persons
    0.84
     adversely
    0.83
     applicable
    0.82
     persons
    0.82
    Activations Density 0.262%

    No Known Activations