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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & 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. 130535
    Prev
    Next
    INDEX
    Explanations

    that

    np_max-act · gemini-2.0-flash

    This neuron activates on the AI’s role‐definition phrase, e.g. “You are an AI who performs…,” identifying where the system message states the AI’s identity and function.

    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
    Interactive
    -0.07
    .lock
    -0.07
    estruction
    -0.06
    Comparison
    -0.06
     publish
    -0.06
     cit
    -0.06
    banana
    -0.06
     Leisure
    -0.06
    Minus
    -0.06
     Ambient
    -0.06
    POSITIVE LOGITS
     उसन
    0.07
    ('
    0.07
    ،
    0.07
     ALOG
    0.07
     [];
    0.07
    (\'
    0.06
    phant
    0.06
     ،
    0.06
    ":{"
    0.06
    基地
    0.06
    Activations Density 0.052%

    No Known Activations