© 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. 1-RES-JB
    11. 23095
    Prev
    Next
    INDEX
    Explanations

    action words related to physical engagement like wrestling, grappling, and clawing

    oai_token-act-pair · gpt-3.5-turbo

    variations of the word "wrestle" and related terms

    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
     [+
    -0.66
    anamo
    -0.62
    liv
    -0.62
     Koen
    -0.61
    /-
    -0.60
    ocl
    -0.58
     Active
    -0.58
    atten
    -0.58
    USE
    -0.58
     Invisible
    -0.57
    POSITIVE LOGITS
     wrest
    1.22
    restling
    1.07
     Wrest
    0.92
     wrestle
    0.89
    iour
    0.81
     wrestling
    0.79
    eering
    0.79
     Jiu
    0.78
    SourceFile
    0.78
    keye
    0.77
    Activations Density 0.006%

    No Known Activations