© 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. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 128677
    Prev
    Next
    INDEX
    Explanations

    to

    np_max-act · gemini-2.0-flash

    The neuron detects words and short phrases where the model refers to its own programming, ethical constraints, or policy‐driven behavior (e.g. “programmed,” “respectful,” “honest,” “ethical,” “follow”).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.11.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     collection
    -0.07
     ROM
    -0.07
     �
    -0.07
     �
    -0.07
    _rating
    -0.07
     Dimit
    -0.06
     Br
    -0.06
     collapses
    -0.06
     protestors
    -0.06
     developer
    -0.06
    POSITIVE LOGITS
    ;?>
    0.07
    ứ
    0.07
    edula
    0.06
    rowse
    0.06
     ẩm
    0.06
     apiUrl
    0.06
    356
    0.06
    ैठक
    0.06
     узн
    0.06
    gis
    0.06
    Activations Density 0.023%

    No Known Activations