© 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. 63368
    Prev
    Next
    INDEX
    Explanations

    mentions of planets and their characteristics, particularly in relation to discoveries in astronomy

    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
    pty
    -0.15
    igor
    -0.15
     Prosper
    -0.14
    horn
    -0.14
    ensch
    -0.14
     reve
    -0.14
    опол
    -0.14
    اÙĨÙĩ
    -0.14
    otal
    -0.14
    atore
    -0.13
    POSITIVE LOGITS
     planet
    0.41
     planets
    0.38
    Planet
    0.33
     Planet
    0.33
     bodies
    0.32
    planet
    0.31
     Bodies
    0.31
     planetary
    0.28
     body
    0.27
     PLAN
    0.27
    Activations Density 0.049%

    No Known Activations