© 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
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 15685
    Prev
    Next
    INDEX
    Explanations

    references to upcoming steps or changes

    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
    åύ
    -0.15
    .central
    -0.15
    heit
    -0.15
    laps
    -0.14
    ospel
    -0.14
    imits
    -0.14
    нÑıми
    -0.13
    ILA
    -0.13
    redirectTo
    -0.13
    spa
    -0.13
    POSITIVE LOGITS
    -door
    0.34
    -generation
    0.31
    door
    0.26
    /current
    0.24
     generation
    0.23
    -gen
    0.22
    -best
    0.21
     few
    0.21
    -big
    0.21
     big
    0.20
    Activations Density 0.056%

    No Known Activations