Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 41996
    Prev
    Next
    INDEX
    Explanations

    Justice/vigilantism

    np_max-act · gemini-2.0-flash

    The neuron primarily detects numeric tokens and numeric-like strings (e.g. numbers, decimals, dates, and other digit-containing or hyphenated numeric expressions).

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    The neuron detects mentions of vigilantism or extrajudicial punishment (words like vigilante, vigil, executioners, calls for death/retribution).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0

    mentions of acting outside official authority—self-appointed roles or enforcement, especially vigilantism and “taking the law into one’s own hands.”

    oai_token-act-pair · gpt-5Triggered by @vetterc0
    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
    か
    -0.07
    SEND
    -0.07
    .unwrap
    -0.07
     THC
    -0.07
    汗
    -0.06
    Sweden
    -0.06
     Fotos
    -0.06
     Luxury
    -0.06
     enquiry
    -0.06
    _mono
    -0.06
    POSITIVE LOGITS
     statically
    0.07
    closing
    0.06
    important
    0.06
    LEGRO
    0.06
     《
    0.06
    :normal
    0.06
     확실
    0.06
     四
    0.06
    ätt
    0.06
    .cells
    0.06
    Activations Density 0.235%

    No Known Activations