© 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. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 10-RES-JB
    11. 4407
    Prev
    Next
    INDEX
    Explanations

    adjectives and nouns describing specific groups of people

    oai_token-act-pair · gpt-3.5-turbo
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    agra
    -0.65
    ãĥ¡
    -0.64
    malink
    -0.61
    inventoryQuantity
    -0.56
    Redditor
    -0.55
    ords
    -0.55
    Pg
    -0.54
    %]
    -0.54
    cision
    -0.54
    ivated
    -0.54
    POSITIVE LOGITS
     how
    1.10
     why
    1.04
     goodbye
    0.97
     that
    0.91
     what
    0.90
     about
    0.85
    that
    0.84
     beware
    0.80
    why
    0.80
     WHY
    0.79
    Activations Density 0.172%

    No Known Activations