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. 84367
    Prev
    Next
    INDEX
    Explanations

    most

    np_max-act · gemini-2.0-flash

    The neuron activates on occurrences of “most” (and its variant “mostly”), i.e. the intensifier or superlative “most.”

    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
    avě
    -0.08
     Bien
    -0.08
     Valentine
    -0.07
     veil
    -0.07
    /xml
    -0.07
     Bag
    -0.07
     увели
    -0.07
     Eagle
    -0.07
     человека
    -0.07
    Bien
    -0.07
    POSITIVE LOGITS
     Most
    0.14
    Most
    0.12
    most
    0.12
     most
    0.11
    -most
    0.10
    MOST
    0.09
    st
    0.09
     MOST
    0.09
    HOST
    0.08
    rst
    0.08
    Activations Density 0.027%

    No Known Activations