© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 24425
    Prev
    Next
    INDEX
    Explanations

    descriptors of demeanor and character, particularly related to politeness and authority

    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
    iasi
    -0.16
    //{{
    -0.16
    atürk
    -0.13
    okino
    -0.13
    aison
    -0.13
    aeda
    -0.13
     Bout
    -0.13
    eyh
    -0.13
    .yy
    -0.13
    774
    -0.12
    POSITIVE LOGITS
     tone
    0.26
     manner
    0.24
     approach
    0.21
     tones
    0.21
     yet
    0.21
     demeanor
    0.20
     nature
    0.18
    ness
    0.18
     smile
    0.17
     delivery
    0.17
    Activations Density 0.190%

    No Known Activations