© 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. 23914
    Prev
    Next
    INDEX
    Explanations

    expressions indicating decline or deterioration

    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
     away
    -0.21
    Away
    -0.17
    away
    -0.16
    اÙĦص
    -0.15
     down
    -0.15
    ault
    -0.15
     Away
    -0.15
    t
    -0.15
    arily
    -0.14
    -away
    -0.14
    POSITIVE LOGITS
    river
    0.20
    hill
    0.19
    stairs
    0.16
     hill
    0.16
     stairs
    0.16
     fighting
    0.15
    swing
    0.15
    Hard
    0.15
    gang
    0.15
    eca
    0.15
    Activations Density 0.030%

    No Known Activations