© 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. 32995
    Prev
    Next
    INDEX
    Explanations

    phrases that include the word "for" followed by a mention of a subject or item

    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
    icable
    -0.15
     somew
    -0.15
    ely
    -0.14
    alone
    -0.14
    .mock
    -0.14
    ima
    -0.14
    ĻĤ
    -0.14
    hle
    -0.14
    iska
    -0.14
    afil
    -0.14
    POSITIVE LOGITS
     that
    0.37
    that
    0.30
     rằng
    0.28
     crying
    0.26
    That
    0.24
     That
    0.24
    	that
    0.22
     že
    0.22
     cry
    0.21
     ÑĩÑĤо
    0.21
    Activations Density 0.022%

    No Known Activations