© 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 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. 42692
    Prev
    Next
    INDEX
    Explanations

    information related to emergency response and evacuation services

    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
    oto
    -0.16
    andid
    -0.14
    çīĩ
    -0.14
    nÃŃ
    -0.13
    omin
    -0.13
    Performed
    -0.13
    รà¸ģ
    -0.13
    LOTS
    -0.13
    guards
    -0.12
    å¤ļãģĦ
    -0.12
    POSITIVE LOGITS
     anyone
    0.37
     any
    0.32
     Anyone
    0.29
     anybody
    0.27
    Anyone
    0.27
    Any
    0.27
     Any
    0.26
     If
    0.26
     if
    0.25
    ä»»ä½ķ
    0.23
    Activations Density 0.124%

    No Known Activations