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. 84696
    Prev
    Next
    INDEX
    Explanations

    "to be" verb forms

    np_max-act · gemini-2.0-flash

    The neuron detects words in phrases expressing that something is wrong or not as expected (e.g. “something is not right”).

    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
    .fromJson
    -0.07
    imit
    -0.07
     commentaire
    -0.06
    .categories
    -0.06
     celé
    -0.06
    わたし
    -0.06
    Orm
    -0.06
    IMIT
    -0.06
    JKLM
    -0.06
    Que
    -0.06
    POSITIVE LOGITS
     doen
    0.08
    aling
    0.07
    /store
    0.07
     Climate
    0.07
     Aless
    0.06
    lider
    0.06
     al
    0.06
     something
    0.06
    full
    0.06
    우스
    0.06
    Activations Density 0.016%

    No Known Activations