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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    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. 2553
    Prev
    Next
    INDEX
    Explanations

    constructions indicating surveillance or observation of individuals

    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
     too
    -0.15
    !:
    -0.13
    ains
    -0.13
     far
    -0.13
    too
    -0.12
    본
    -0.12
    691
    -0.12
     const
    -0.12
     Too
    -0.12
    esis
    -0.12
    POSITIVE LOGITS
    /Runtime
    0.16
    linkplain
    0.16
    ÑģÑĤе
    0.15
    ãĥ«ãĥĪ
    0.15
    zon
    0.15
    erece
    0.14
     Narr
    0.14
    ontrol
    0.14
    ãģĵãģĨ
    0.14
    mour
    0.14
    Activations Density 0.012%

    No Known Activations