© 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. 43865
    Prev
    Next
    INDEX
    Explanations

    references to character traits or qualities

    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
    eson
    -0.17
    aja
    -0.16
    halt
    -0.16
    inar
    -0.16
    ëĵĿ
    -0.16
    739
    -0.16
    batim
    -0.16
    ese
    -0.15
    illon
    -0.15
    ali
    -0.15
    POSITIVE LOGITS
     Fi
    0.24
     fi
    0.23
    Fi
    0.21
    -fi
    0.19
    qh
    0.19
    -Fi
    0.18
    rst
    0.18
    CTION
    0.17
    end
    0.17
    endcode
    0.16
    Activations Density 0.003%

    No Known Activations