Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 63170
    Prev
    Next
    INDEX
    Explanations

    references to creative concepts and ideas

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    words or phrases related to concepts, thoughts, or abstract notions, particularly focusing on the word "ideas".

    oai_token-act-pair · claude-3-5-sonnet-20240620Triggered by @aml7990
    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
    istrovstvÃŃ
    -0.17
    endon
    -0.17
    ìĦľëĬĶ
    -0.17
    ello
    -0.16
    nemonic
    -0.16
    plies
    -0.15
    ese
    -0.15
    ir
    -0.14
    weg
    -0.14
    ëĭĺìĿ´
    -0.14
    POSITIVE LOGITS
     behind
    0.17
    beh
    0.17
    oppins
    0.15
    /app
    0.15
    quel
    0.15
     about
    0.14
     starters
    0.14
    732
    0.14
    613
    0.14
    hare
    0.14
    Activations Density 0.036%

    No Known Activations