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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    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. 31380
    Prev
    Next
    INDEX
    Explanations

    phrases related to wealth and its distribution in society

    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
    >null
    -0.15
    thane
    -0.14
    hammer
    -0.14
     Weston
    -0.14
     ÐĿаÑģеление
    -0.13
     expansive
    -0.13
    rov
    -0.13
    icip
    -0.13
    acher
    -0.13
    irim
    -0.13
    POSITIVE LOGITS
     concentration
    0.80
     Concent
    0.76
     concentrations
    0.66
    éĽĨä¸Ń
    0.63
     concentr
    0.60
     concentrated
    0.58
     конÑĨенÑĤÑĢа
    0.50
     concentrate
    0.50
     concentrating
    0.49
     táºŃp
    0.39
    Activations Density 0.036%

    No Known Activations