Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 23761
    Prev
    Next
    INDEX
    Explanations

    verbs indicating states of being or conditions

    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
    ÙħÛĮÙĨ
    -0.15
     Bias
    -0.14
    iden
    -0.13
    ãģĿãģ®ä»ĸ
    -0.13
     predomin
    -0.13
    upert
    -0.13
    _DEPEND
    -0.13
    iffer
    -0.13
    apper
    -0.13
    ereum
    -0.13
    POSITIVE LOGITS
     associated
    0.58
    associated
    0.49
     assoc
    0.44
     Associated
    0.43
    Associated
    0.41
     association
    0.40
     associate
    0.40
     linked
    0.38
     asoci
    0.38
     associ
    0.38
    Activations Density 0.078%

    No Known Activations