© 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. 8761
    Prev
    Next
    INDEX
    Explanations

    punctuation marks and question indicators in the text

    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
     Ðŀдна
    -0.18
    chal
    -0.14
    ÑĢаÑĤи
    -0.14
    âĢŀM
    -0.13
     ÐŃÑĤи
    -0.13
     dÅĻÃŃ
    -0.13
    ehir
    -0.13
    éĤ£éĩĮ
    -0.13
    ÅĻet
    -0.13
    æ§
    -0.13
    POSITIVE LOGITS
    onian
    0.15
    stell
    0.14
    esto
    0.14
     cred
    0.14
    onic
    0.13
    Result
    0.13
    indr
    0.13
    utto
    0.13
     automát
    0.13
    illis
    0.13
    Activations Density 0.016%

    No Known Activations