© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 21164
    Prev
    Next
    INDEX
    Explanations

    phrases that express opinion or assessment

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Negative Logits
    ÃĹ↵↵
    -0.18
    ê¸
    -0.14
    angen
    -0.14
    _Statics
    -0.14
    atern
    -0.14
    illions
    -0.14
    ambiguous
    -0.13
     sounding
    -0.13
    ÃĶNG
    -0.13
    icari
    -0.13
    POSITIVE LOGITS
     says
    0.54
     speaks
    0.46
     say
    0.45
     Says
    0.45
     speak
    0.42
     tells
    0.40
     tell
    0.39
    says
    0.38
    say
    0.33
     Say
    0.32
    Activations Density 0.071%

    No Known Activations