© 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 EvalsExportsinterp-engineNEWAPI 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. 11-RES-JB
    11. 12209
    Prev
    Next
    INDEX
    Explanations

    locations or places - whether they are cities, countries, or other specific geographical areas

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "other" and related quantifiers

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    opolis
    -0.53
    Â
    -0.50
    another
    -0.50
    let
    -0.49
    achment
    -0.49
    pper
    -0.49
    cture
    -0.49
    rium
    -0.49
    pping
    -0.49
    emaker
    -0.49
    POSITIVE LOGITS
     subreddits
    0.87
     embodiments
    0.82
     poems
    0.79
     recipes
    0.78
     galaxies
    0.75
     languages
    0.73
     patents
    0.71
     actresses
    0.70
     breweries
    0.69
     cancers
    0.69
    Activations Density 0.673%

    No Known Activations