© 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. 41801
    Prev
    Next
    INDEX
    Explanations

    references to the word "Tiger" and its variations

    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
    abilit
    -0.17
    agma
    -0.17
    soever
    -0.16
    tridge
    -0.16
    zione
    -0.16
    tes
    -0.15
    vest
    -0.15
    exampleModal
    -0.15
    ibold
    -0.15
    tesy
    -0.15
    POSITIVE LOGITS
     Woods
    0.26
     cub
    0.24
    hawk
    0.21
     claw
    0.21
     Cubs
    0.20
     Claw
    0.20
     Tiger
    0.20
     Cub
    0.20
     Lily
    0.18
    -striped
    0.18
    Activations Density 0.007%

    No Known Activations