© 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. 41076
    Prev
    Next
    INDEX
    Explanations

    compound adjectives and phrases that indicate a specific orientation or focus

    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
     latter
    -0.24
    /her
    -0.22
    âĢIJ
    -0.16
    acades
    -0.15
     तरह
    -0.14
    agua
    -0.14
    chop
    -0.14
    ses
    -0.13
    æĸ°çļĦ
    -0.13
    -level
    -0.13
    POSITIVE LOGITS
    /-
    0.32
    ness
    0.19
    adays
    0.15
    ned
    0.15
    ÂĿ
    0.15
    enticate
    0.15
    NESS
    0.15
    enticator
    0.15
    edly
    0.14
    /version
    0.14
    Activations Density 0.510%

    No Known Activations