© 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. 14194
    Prev
    Next
    INDEX
    Explanations

    words related to intense or forceful actions

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the term "thr" and its variations, possibly indicating focus on thrill or excitement

    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
    enegger
    -0.77
     lying
    -0.68
     pred
    -0.68
    DEN
    -0.65
     suited
    -0.65
     prone
    -0.64
     bed
    -0.64
     kneeling
    -0.63
     posed
    -0.57
     beds
    -0.57
    POSITIVE LOGITS
    ifty
    1.46
    ongs
    1.39
    ashing
    1.36
    uster
    1.33
    ift
    1.31
    usters
    1.31
    ills
    1.23
    ashed
    1.22
    ice
    1.20
    ifted
    1.18
    Activations Density 0.021%

    No Known Activations