© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 43169
    Prev
    Next
    INDEX
    Explanations

    phrases related to promises and assertions made by political figures

    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
    upal
    -0.16
    æ¯
    -0.15
    è¤ĩ
    -0.15
     seins
    -0.15
    WISE
    -0.15
    XE
    -0.15
    кÑĤÑĥ
    -0.14
    ZW
    -0.14
     Zeit
    -0.14
    Ĭ¶
    -0.14
    POSITIVE LOGITS
     Melania
    0.25
     Drain
    0.23
    177
    0.19
     Bannon
    0.19
     America
    0.18
     Mattis
    0.18
     American
    0.18
     MAG
    0.17
     drain
    0.17
     Trump
    0.17
    Activations Density 0.077%

    No Known Activations