© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 15565
    Prev
    Next
    INDEX
    Explanations

    prepositions and connecting words that show relationships between ideas

    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
    822
    -0.14
    acter
    -0.14
    ENCHMARK
    -0.14
    ItemAt
    -0.14
    /or
    -0.14
    umph
    -0.14
    imity
    -0.13
    ragon
    -0.13
    IJ
    -0.13
    ाष
    -0.13
    POSITIVE LOGITS
    ingly
    0.20
    ously
    0.20
     fois
    0.20
    -wise
    0.19
    wise
    0.19
    itionally
    0.18
    ifr
    0.17
    ually
    0.17
    aneously
    0.17
    beit
    0.17
    Activations Density 0.024%

    No Known Activations