© 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. 30152
    Prev
    Next
    INDEX
    Explanations

    instructions and mechanisms for gameplay in a Pokémon context

    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
    èľĺèĽĽ
    -0.20
    ante
    -0.16
    gaard
    -0.16
    ocs
    -0.15
     Panther
    -0.14
     Oxford
    -0.14
    agle
    -0.14
     Zot
    -0.14
     Weaver
    -0.14
     Manit
    -0.14
    POSITIVE LOGITS
     Pokémon
    0.36
     Pok
    0.33
    Pok
    0.31
     Trainer
    0.31
     trainers
    0.31
     trainer
    0.30
     Pokemon
    0.29
    trainer
    0.29
     pok
    0.27
     pokemon
    0.26
    Activations Density 0.067%

    No Known Activations