© 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
    3. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 29960
    Prev
    Next
    INDEX
    Explanations

    possessive forms and phrases indicating ownership or relation

    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
     Shoulder
    -0.17
     Neck
    -0.17
    neck
    -0.15
    -neck
    -0.15
    .ov
    -0.15
     shoulder
    -0.15
     Chess
    -0.15
    ">//
    -0.14
     cervical
    -0.14
     neck
    -0.14
    POSITIVE LOGITS
     foot
    0.73
     feet
    0.70
     Foot
    0.65
    foot
    0.62
     Feet
    0.60
    Foot
    0.60
     footwear
    0.58
    -foot
    0.57
     FOOT
    0.51
     toes
    0.50
    Activations Density 0.023%

    No Known Activations