© 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. 43755
    Prev
    Next
    INDEX
    Explanations

    information regarding the safety and effects of chemicals in farm-raised fish

    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
    avar
    -0.17
    angi
    -0.16
    vrier
    -0.15
    nesia
    -0.15
    StackNavigator
    -0.15
    cth
    -0.14
    bsd
    -0.14
    IVEN
    -0.14
    .IsNullOr
    -0.14
    Ùħد
    -0.14
    POSITIVE LOGITS
     vitamin
    0.43
     Vitamin
    0.40
    vit
    0.37
     Vit
    0.31
     vitam
    0.28
     vit
    0.26
    itamin
    0.24
     vitamins
    0.24
     sunshine
    0.22
     ret
    0.21
    Activations Density 0.004%

    No Known Activations