© 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. 54796
    Prev
    Next
    INDEX
    Explanations

    first-person pronouns and expressions of personal reflection or identity

    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
     indeed
    -0.18
    lder
    -0.15
     appar
    -0.15
    ussed
    -0.14
     Indeed
    -0.14
    duc
    -0.14
    íĥĪ
    -0.14
    rror
    -0.14
     actually
    -0.14
    lds
    -0.14
    POSITIVE LOGITS
     wondering
    0.19
     suddenly
    0.19
     STILL
    0.16
    still
    0.16
     yet
    0.16
     Still
    0.15
    alic
    0.15
    itori
    0.15
    еÑī
    0.15
     therefore
    0.15
    Activations Density 0.013%

    No Known Activations