© 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. 12315
    Prev
    Next
    INDEX
    Explanations

    expressions that introduce alternative possibilities or questions

    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
    icorn
    -0.16
    ìĿ´íĦ°
    -0.14
    &amp
    -0.14
     NONINFRINGEMENT
    -0.14
    NavController
    -0.14
    ongs
    -0.13
    à¸Ĥ
    -0.13
    ÄIJT
    -0.13
    orks
    -0.13
    ÂĿ
    -0.13
    POSITIVE LOGITS
     perhaps
    0.36
     maybe
    0.35
     even
    0.34
     else
    0.33
     indeed
    0.29
     worse
    0.28
    maybe
    0.27
    phan
    0.27
    perhaps
    0.25
     equival
    0.25
    Activations Density 0.148%

    No Known Activations