© 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. 43154
    Prev
    Next
    INDEX
    Explanations

    punctuation marks indicating significant points or endings in 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
    _SUITE
    -0.15
    )prepare
    -0.14
    laps
    -0.14
    agma
    -0.14
    _flg
    -0.14
    arest
    -0.13
    usta
    -0.13
    wij
    -0.13
    essim
    -0.13
     weer
    -0.13
    POSITIVE LOGITS
     second
    1.01
     Second
    0.92
    Second
    0.92
    second
    0.88
     Secondly
    0.87
     SECOND
    0.77
    -second
    0.75
    第äºĮ
    0.75
    _second
    0.71
    .second
    0.70
    Activations Density 0.033%

    No Known Activations