© 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. 3-RES-JB
    11. 15934
    Prev
    Next
    INDEX
    Explanations

    words related to Major League sports, particularly baseball and soccer

    oai_token-act-pair · gpt-3.5-turbo

    references to professional sports leagues, particularly Major League Baseball and Major League Soccer

    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
    ãĤ¨ãĥ«
    -0.83
    atche
    -0.79
    =~=~
    -0.79
    ivated
    -0.72
    PsyNetMessage
    -0.69
    yz
    -0.68
    EStream
    -0.67
    reve
    -0.64
     sidx
    -0.64
    ãĥ¼ãĤ¯
    -0.63
    POSITIVE LOGITS
     Baseball
    1.03
     Soccer
    0.89
     pitcher
    0.76
     Athletics
    0.72
    hander
    0.71
    League
    0.70
     Athlet
    0.70
    cano
    0.69
    ructose
    0.69
     stadiums
    0.69
    Activations Density 0.035%

    No Known Activations