© 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. 19988
    Prev
    Next
    INDEX
    Explanations

    items or elements that are part of a larger group

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "few" in various contexts

    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
    acus
    -0.97
    UAL
    -0.78
    akings
    -0.72
    ller
    -0.69
    alam
    -0.69
    amen
    -0.68
    roit
    -0.68
    abella
    -0.67
    aucas
    -0.67
    ilon
    -0.66
    POSITIVE LOGITS
     dozen
    1.26
     hundred
    1.11
     thousand
    0.94
     weeks
    0.93
    dozen
    0.90
     nights
    0.86
    een
    0.82
     pointers
    0.82
     days
    0.80
     handy
    0.80
    Activations Density 0.034%

    No Known Activations