© 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. 43357
    Prev
    Next
    INDEX
    Explanations

    references to prior occurrences or states

    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
    ointed
    -0.17
    llib
    -0.15
    anje
    -0.15
    olla
    -0.15
    ãĥ¼ãĥ¬
    -0.15
    ane
    -0.15
    ened
    -0.15
    usan
    -0.15
    nbsp
    -0.15
    ToWorld
    -0.15
    POSITIVE LOGITS
     unknown
    0.27
    unknown
    0.26
     known
    0.22
     thought
    0.22
    Unknown
    0.20
     never
    0.19
    _UNKNOWN
    0.19
     unheard
    0.18
     unseen
    0.18
     held
    0.17
    Activations Density 0.034%

    No Known Activations