© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Llama3.3-70B-IT
    5. 50-RESID-POST-GF
    7. 49309
    Prev
    Next
    INDEX
    Explanations

    numbers separated by newlines

    np_acts-logits-general · gemini-2.5-flash-lite

    structural or formatting tokens—such as special header markers, tabs, and punctuation/symbol characters—indicating document or code-like structure rather than regular words.

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With LLAMA3.3-70B-IT @ 50-resid-post-gf
    Configuration
    Goodfire/Llama-3.3-70B-Instruct-SAE-l50/Llama-3.3-70B-Instruct-SAE-l50.pt
    Prompts (Dashboard)
    10,000 prompts, 128 tokens each
    Dataset (Dashboard)
    lmsys/lmsys-chat-1m
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     IsNot
    -0.08
    .GraphicsUnit
    -0.08
    atürk
    -0.07
    æ¢
    -0.07
    Brains
    -0.07
    iв
    -0.07
    SystemService
    -0.07
    icode
    -0.07
     ฿
    -0.07
    rollo
    -0.07
    POSITIVE LOGITS
    ¦æĥħ
    0.11
    /goto
    0.08
    è¾ij
    0.08
     Garner
    0.07
    uli
    0.07
     neod
    0.07
     Hod
    0.07
     redundant
    0.07
    lich
    0.07
    ©©
    0.07
    Activations Density 0.516%

    No Known Activations