© 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
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 24028
    Prev
    Next
    INDEX
    Explanations

    words indicating personal pronouns, particularly "you" and "them"

    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
     of
    -0.20
     all
    -0.17
     ÏĦηÏĤ
    -0.15
     both
    -0.14
    ;amp
    -0.14
     вÑģеÑħ
    -0.14
     and
    -0.13
     themselves
    -0.13
    ’t
    -0.13
     ÑĤобÑĤо
    -0.13
    POSITIVE LOGITS
    /us
    0.21
    /her
    0.19
    zelf
    0.18
     are
    0.18
    ##_
    0.17
    -même
    0.16
    /th
    0.16
     ÙħÛĮÙĦادÛĮ
    0.16
    curity
    0.15
    -ci
    0.14
    Activations Density 0.087%

    No Known Activations