© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 3-RES-JB
    11. 17060
    Prev
    Next
    INDEX
    Explanations

    words related to a sudden and intense attack or promotion

    oai_token-act-pair · gpt-3.5-turbo

    references to "blitz," particularly in a context related to intense activities or attacks

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Negative Logits
    ndra
    -0.78
    AFTA
    -0.77
    inguishable
    -0.73
    uct
    -0.71
     Norn
    -0.70
    uana
    -0.70
    rouch
    -0.70
    akin
    -0.69
    aft
    -0.68
    avery
    -0.67
    POSITIVE LOGITS
     blitz
    1.24
    rieg
    0.96
    ãĥĦ
    0.91
    schild
    0.85
    éĹĺ
    0.84
     Blitz
    0.82
    naires
    0.81
    enegger
    0.80
    entimes
    0.80
    iness
    0.79
    Activations Density 0.007%

    No Known Activations