© 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. 24462
    Prev
    Next
    INDEX
    Explanations

    references to social structures and systems of classification

    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
     along
    -0.17
    ãģĿãģĹãģ¦
    -0.16
    along
    -0.15
     alongside
    -0.15
     together
    -0.15
    atra
    -0.15
    AND
    -0.14
     además
    -0.14
    以åıĬ
    -0.13
    Ñĥки
    -0.13
    POSITIVE LOGITS
     váºŃy
    0.28
     except
    0.26
     where
    0.23
     ÙħØ«ÙĦا
    0.23
     example
    0.23
     Except
    0.23
     whose
    0.23
     or
    0.22
    whose
    0.21
     oder
    0.21
    Activations Density 0.125%

    No Known Activations