© 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. 41881
    Prev
    Next
    INDEX
    Explanations

    comparative and relational phrases that describe physical locations or scenarios

    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
    unn
    -0.14
    metis
    -0.14
    âĶIJ
    -0.13
    mtree
    -0.13
     Locations
    -0.13
    aea
    -0.13
    atif
    -0.13
    .locations
    -0.13
    ãģŀ
    -0.12
    vn
    -0.12
    POSITIVE LOGITS
     nearby
    0.83
     surrounding
    0.76
     neighboring
    0.70
     adjacent
    0.68
     neighbouring
    0.60
     Nearby
    0.48
     surround
    0.47
     Surround
    0.47
    adj
    0.46
     adjoining
    0.46
    Activations Density 0.063%

    No Known Activations