© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    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. 90729
    Prev
    Next
    INDEX
    Explanations

    locations

    np_max-act · gemini-2.0-flash

    This neuron detects multiword proper names that pair a capitalized geographic or organizational identifier with a generic class noun (e.g. “Field Office,” “Fuel Station,” “Herd Management Area,” “Fishery,” “Commandery,” “Landfill,” “Craton”).

    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
     Marcus
    -0.07
     verir
    -0.07
    \web
    -0.07
     AK
    -0.06
     Dix
    -0.06
    (block
    -0.06
    :^{↵
    -0.06
    ascade
    -0.06
    ffective
    -0.06
    せる
    -0.06
    POSITIVE LOGITS
    �
    0.07
    			       
    0.07
    ращения
    0.06
    Au
    0.06
     hug
    0.06
     الاتحاد
    0.06
    tif
    0.06
    นๆ
    0.06
     Similarly
    0.06
    .getUrl
    0.06
    Activations Density 0.081%

    No Known Activations