© 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. 3-RES-JB
    11. 9462
    Prev
    Next
    INDEX
    Explanations

    the word "surprise" with a high relevance

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the word "surprise."

    oai_token-act-pair · gpt-4o-miniTriggered 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
    Not in Any Lists

    No Comments

    Negative Logits
    ©¶æ
    -1.08
    İĭ
    -0.86
    nan
    -0.79
    everal
    -0.77
    bis
    -0.75
    ĪĴ
    -0.74
    Ķ
    -0.73
    burgh
    -0.72
    ccording
    -0.72
    lim
    -0.69
    POSITIVE LOGITS
     Surprise
    0.98
     surprise
    0.97
     surprises
    0.93
    ingly
    0.79
     coincidence
    0.72
    ously
    0.70
    Berry
    0.69
     onlook
    0.68
     Pon
    0.68
    imaru
    0.67
    Activations Density 0.019%

    No Known Activations