© 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. 23016
    Prev
    Next
    INDEX
    Explanations

    geographic terms and descriptions related to regions and rivers

    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
    reau
    -0.16
    ÙĩÙĨÚ¯
    -0.15
    unca
    -0.15
    ìĽĶ
    -0.14
    íĶ½
    -0.14
     côt
    -0.14
    iker
    -0.14
    ÑģÑĤÑĢи
    -0.14
    .CG
    -0.14
    fsp
    -0.14
    POSITIVE LOGITS
     pr
    0.48
     Midwest
    0.46
     Mid
    0.38
     Prairie
    0.37
     Kansas
    0.36
     Great
    0.36
     Pra
    0.36
    Mid
    0.35
     Missouri
    0.33
     Nebraska
    0.32
    Activations Density 0.055%

    No Known Activations