© 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. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 12872
    Prev
    Next
    INDEX
    Explanations

    expressions of excitement or emphasis

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    gated
    Context Size
    1,024
    Dataset
    Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    rb
    -0.16
    inness
    -0.14
    ampler
    -0.14
    баÑĩ
    -0.14
    getOption
    -0.14
    amp
    -0.14
    specialchars
    -0.13
    inity
    -0.13
    iring
    -0.13
    ony
    -0.13
    POSITIVE LOGITS
     also
    0.24
     wait
    0.23
    Also
    0.21
    also
    0.21
     AND
    0.20
     incident
    0.20
     speaking
    0.20
     and
    0.20
     And
    0.20
     à¹ģละ
    0.19
    Activations Density 0.006%

    No Known Activations