© 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. 25525
    Prev
    Next
    INDEX
    Explanations

    getting caught or tricked

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     annihil
    -0.11
     åŀ
    -0.09
     vending
    -0.09
     scav
    -0.09
    /www
    -0.09
     inflict
    -0.09
    destruct
    -0.08
    BootApplication
    -0.08
    AtA
    -0.08
    ---------\n\n
    -0.08
    POSITIVE LOGITS
     harm
    0.20
     attack
    0.17
     kill
    0.16
     harmed
    0.15
     Harm
    0.14
     killed
    0.14
     mist
    0.13
     Attack
    0.12
     capture
    0.12
     kills
    0.12
    Activations Density 0.070%

    No Known Activations