© 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. 8-RES-JB
    11. 24540
    Prev
    Next
    INDEX
    Explanations

    instances where a phrase begins with "Oh, and" followed by additional information or action

    oai_token-act-pair · gpt-3.5-turbo

    expressions of surprise or exclamation

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Ö¼
    -0.79
    actionDate
    -0.78
    BILITIES
    -0.77
    alion
    -0.73
    orea
    -0.72
    arov
    -0.71
    arij
    -0.71
    emo
    -0.71
    omething
    -0.70
    arcity
    -0.68
    POSITIVE LOGITS
     heavens
    0.70
     dear
    0.70
     Merlin
    0.64
     virginity
    0.63
     pardon
    0.63
     surprise
    0.62
     bless
    0.61
     Dynamo
    0.61
     Despair
    0.60
     Miss
    0.60
    Activations Density 0.043%

    No Known Activations