© 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. 44119
    Prev
    Next
    INDEX
    Explanations

    phrases related to the concept of presence or location

    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
    ting
    -0.16
    reste
    -0.16
    .gs
    -0.15
    andin
    -0.15
    vb
    -0.15
     ÑĤам
    -0.15
    eliness
    -0.14
    ãģĬãĤĬ
    -0.14
    cing
    -0.14
    rol
    -0.14
    POSITIVE LOGITS
    abouts
    0.41
    after
    0.34
    ina
    0.27
    jÅ¡ÃŃ
    0.24
    upon
    0.22
    under
    0.22
    -même
    0.21
    ìłĢ
    0.20
    unto
    0.19
    ifr
    0.17
    Activations Density 0.087%

    No Known Activations