© 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. 24905
    Prev
    Next
    INDEX
    Explanations

    phrases indicating difficulty or challenges

    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
    bsolute
    -0.17
    à¤Łà¤ķ
    -0.15
    rors
    -0.15
    oret
    -0.15
    mey
    -0.14
    ãĤĵãģ¨
    -0.14
    ạ
    -0.14
    Stmt
    -0.14
    oretical
    -0.14
    HTTPRequestOperation
    -0.14
    POSITIVE LOGITS
     not
    0.24
     NOT
    0.24
    wig
    0.18
    not
    0.17
    ESC
    0.16
    Not
    0.16
     Yak
    0.16
    ä¸įåĪ°
    0.15
     Not
    0.15
    .fm
    0.15
    Activations Density 0.014%

    No Known Activations