© 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. 42902
    Prev
    Next
    INDEX
    Explanations

    phrases indicating increases or enhancements

    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
    æĽ´å¤ļ
    -0.17
     otherwise
    -0.17
     more
    -0.16
     anymore
    -0.15
    smart
    -0.14
    remely
    -0.14
    ignum
    -0.14
    ãĤĤãģĨ
    -0.14
    ptom
    -0.14
     mais
    -0.14
    POSITIVE LOGITS
    -than
    0.32
     than
    0.32
    than
    0.26
    -or
    0.20
    Than
    0.20
     THAN
    0.19
    tti
    0.19
     Than
    0.18
     än
    0.18
    zzarella
    0.17
    Activations Density 0.107%

    No Known Activations