© 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. 29558
    Prev
    Next
    INDEX
    Explanations

    punctuation marks and formatting symbols typically associated with lists or data structures

    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
    oval
    -0.16
    Jackson
    -0.15
     ủy
    -0.15
     Emil
    -0.15
     Shaft
    -0.14
    .setPrototypeOf
    -0.14
    สำà¸Ļ
    -0.14
    43
    -0.14
    65
    -0.14
    лÑĸн
    -0.14
    POSITIVE LOGITS
    300
    0.48
    299
    0.43
    298
    0.39
    301
    0.39
    302
    0.37
    297
    0.37
    304
    0.36
    296
    0.35
    303
    0.35
    290
    0.34
    Activations Density 0.015%

    No Known Activations