© 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. OpenMOSS · Llama Scope: SAEs for Llama-3.1-8B
    5. Llama3.1-8B (Base)
    7. Residual Stream
    9. 17-LLAMASCOPE-RES-131K
    11. 95791
    Prev
    Next
    INDEX
    Explanations

    instances of the word "pretend" and its variations, indicating themes of deception or role-playing

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    cerebras/SlimPajama-627B
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    erk
    -0.07
    ader
    -0.07
    adia
    -0.07
    lets
    -0.06
    dump
    -0.06
    ackers
    -0.06
    hood
    -0.06
     freopen
    -0.06
    orges
    -0.06
    rip
    -0.06
    POSITIVE LOGITS
    inkle
    0.08
    entially
    0.08
     rằng
    0.07
    ÛĮز
    0.07
    /assert
    0.07
    mÄĽ
    0.07
    atively
    0.07
    ENT
    0.07
    ECTOR
    0.07
     Bender
    0.06
    Activations Density 0.003%

    No Known Activations