© 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. 12899
    Prev
    Next
    INDEX
    Explanations

    specific measurements and characteristics of animals

    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
    .scalablytyped
    -0.18
     lizard
    -0.17
     Lamb
    -0.17
    mlin
    -0.17
     Duck
    -0.16
     shark
    -0.16
    à¸Ħà¸Ļ
    -0.15
     Dolphin
    -0.14
     Falcon
    -0.14
     Dog
    -0.14
    POSITIVE LOGITS
     adult
    0.33
     male
    0.29
     individuals
    0.28
     males
    0.28
     female
    0.27
     females
    0.27
    male
    0.24
     adults
    0.23
     mature
    0.23
     Adult
    0.23
    Activations Density 0.061%

    No Known Activations