© 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. 29051
    Prev
    Next
    INDEX
    Explanations

    punctuation and phrases indicating continuation or elaboration

    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
     desar
    -0.15
    oldem
    -0.14
    dit
    -0.14
    ingga
    -0.13
    ashi
    -0.13
    _documento
    -0.13
    ê¸Ķ
    -0.13
    but
    -0.13
    erved
    -0.13
     à¹Ģà¸ŀราะ
    -0.13
    POSITIVE LOGITS
     which
    0.31
    which
    0.24
     though
    0.23
     however
    0.23
     like
    0.23
     if
    0.23
     while
    0.20
    once
    0.19
     Which
    0.19
    Which
    0.19
    Activations Density 0.495%

    No Known Activations