© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 43474
    Prev
    Next
    INDEX
    Explanations

    instances of the word "to" and its variations indicating purpose or intention

    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
     ampl
    -0.17
    ember
    -0.16
    zza
    -0.15
    ASURE
    -0.15
    jah
    -0.14
    iber
    -0.14
    ustain
    -0.14
    ç°¡
    -0.14
    cribe
    -0.14
    Compute
    -0.14
    POSITIVE LOGITS
     ensure
    0.23
     help
    0.19
     enable
    0.17
     ermög
    0.17
    help
    0.17
     guarantee
    0.16
     attempt
    0.16
    allow
    0.16
     increase
    0.16
     ensures
    0.15
    Activations Density 0.097%

    No Known Activations