© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 4-RES-JB
    11. 12114
    Prev
    Next
    INDEX
    Explanations

    numeric values related to medical units of measurement

    oai_token-act-pair · gpt-3.5-turbo

    quantities and measurements related to substances and their concentrations

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 4-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    4
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ļéĨĴ
    -0.73
    Redditor
    -0.73
    FORE
    -0.71
    FTWARE
    -0.68
     Inquisition
    -0.67
    ESCO
    -0.65
    Catalog
    -0.62
     jan
    -0.61
    atural
    -0.60
     Forge
    -0.60
    POSITIVE LOGITS
    mt
    1.03
    mg
    0.89
    omez
    0.81
    ERSON
    0.81
    iants
    0.81
     dose
    0.81
    onz
    0.79
    TPS
    0.79
    rowth
    0.78
    allery
    0.78
    Activations Density 0.008%

    No Known Activations