© 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. 26679
    Prev
    Next
    INDEX
    Explanations

    numerical values and associated statistics

    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
    373
    +0.30
    1.6%
    87
    +0.14
    0.7%
    480
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    430
    +0.30
    0.10
    496
    +0.14
    0.07
    316
    +0.10
    0.09
    Negative Logits
     flyers
    -0.58
     pressuring
    -0.57
     transc
    -0.56
     horizont
    -0.56
     univers
    -0.55
     carrot
    -0.55
     codes
    -0.53
     consistency
    -0.52
     plagiar
    -0.52
     motto
    -0.51
    POSITIVE LOGITS
     09
    0.94
     49
    0.91
     39
    0.89
     00
    0.88
     37
    0.88
     34
    0.87
     08
    0.87
     59
    0.86
     33
    0.86
     38
    0.86
    Activations Density 0.781%

    No Known Activations