© 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. 28613
    Prev
    Next
    INDEX
    Explanations

    references to resources and external links

    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
    +0.15
    0.7%
    306
    +0.09
    0.4%
    424
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    33
    +0.15
    0.04
    172
    +0.09
    0.04
    424
    +0.09
    0.04
    Negative Logits
    esson
    -0.65
    ettes
    -0.58
     accur
    -0.56
     horr
    -0.56
    ette
    -0.56
    emo
    -0.55
     cosmetic
    -0.54
    aeda
    -0.54
    actionDate
    -0.54
     lipstick
    -0.53
    POSITIVE LOGITS
     our
    0.88
     www
    0.86
     http
    0.84
     https
    0.83
     Appendix
    0.82
     CONTR
    0.80
     Understanding
    0.78
     Consortium
    0.77
     Getting
    0.74
     Managing
    0.73
    Activations Density 0.623%

    No Known Activations