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

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 84404
    Prev
    Next
    INDEX
    Explanations

    take

    np_max-act · gemini-2.0-flash

    This neuron detects the phrase “Take a” (i.e. suggestions beginning with the verb “take” followed by the article “a”).

    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
    صال
    -0.07
     Operations
    -0.06
     Archbishop
    -0.06
    adders
    -0.06
     mange
    -0.06
    ynchronously
    -0.06
     pH
    -0.06
    орм
    -0.06
     Winchester
    -0.06
    EDIATEK
    -0.06
    POSITIVE LOGITS
     take
    0.08
     taking
    0.08
     Take
    0.08
     Taking
    0.08
    .sun
    0.07
    Taking
    0.07
     Give
    0.07
    	Get
    0.07
    .IMAGE
    0.07
     extra
    0.06
    Activations Density 0.030%

    No Known Activations