© 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
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 41904
    Prev
    Next
    INDEX
    Explanations

    phrases indicating familiarity or knowledge of ongoing discussions or matters

    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
    ëī´ìĬ¤
    -0.14
    zione
    -0.14
     же
    -0.14
    iyas
    -0.13
    inet
    -0.13
    usp
    -0.13
    á»ĵ
    -0.13
    _ENDIAN
    -0.13
    porte
    -0.12
     tslib
    -0.12
    POSITIVE LOGITS
     situation
    0.34
     Situation
    0.29
     matter
    0.29
     deliber
    0.24
     inner
    0.24
     conversations
    0.23
     discussions
    0.23
    matter
    0.22
     proceedings
    0.22
     Matter
    0.22
    Activations Density 0.024%

    No Known Activations