© 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. 19611
    Prev
    Next
    INDEX
    Explanations

    but assistant

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

    the beginning of AI assistant responses, particularly after special formatting tokens.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @yooniel31
    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
    oyer
    -0.10
     Erg
    -0.10
     obvious
    -0.09
    gary
    -0.09
     answered
    -0.09
     Evet
    -0.09
     Answer
    -0.09
     answers
    -0.09
     questions
    -0.09
     answer
    -0.08
    POSITIVE LOGITS
     afraid
    0.10
     Herm
    0.09
    ialog
    0.09
    roe
    0.09
    nnen
    0.09
     Um
    0.08
    åĶ¯
    0.08
     hum
    0.08
    agini
    0.08
    âĶĤ
    0.08
    Activations Density 0.160%

    No Known Activations