© 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 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. 1-RES-JB
    11. 23983
    Prev
    Next
    INDEX
    Explanations

    words with the prefix "rid" followed by a number indicating the activation strength

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "ridiculous."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     largeDownload
    -0.75
     whales
    -0.68
    AU
    -0.66
     Prospect
    -0.62
    ¬¼
    -0.60
     Markets
    -0.59
     Eagle
    -0.58
    SUP
    -0.58
    ULT
    -0.56
     HM
    -0.56
    POSITIVE LOGITS
    rid
    1.18
    gew
    1.03
    gements
    1.02
    anus
    1.00
    gment
    0.94
    ifiers
    0.93
    gments
    0.93
    gement
    0.90
    ged
    0.87
    nesses
    0.85
    Activations Density 0.003%

    No Known Activations