© 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
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 30641
    Prev
    Next
    INDEX
    Explanations

    to be

    np_max-act · gemini-2.0-flash

    This neuron detects comparative or evaluative words and phrases (e.g. “better,” “more effective,” “well‐suited,” “making,” “has,” “also”) that qualify or emphasize a product’s performance or qualities.

    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
     twink
    -0.06
     wakeup
    -0.06
     pitfalls
    -0.06
     minced
    -0.06
     Duel
    -0.05
     HIS
    -0.05
     く
    -0.05
     Reaper
    -0.05
     دختر
    -0.05
     iler
    -0.05
    POSITIVE LOGITS
    服务器
    0.08
                                   
    0.07
    ustrial
    0.07
     Croat
    0.07
     документа
    0.07
    Repository
    0.06
     localVar
    0.06
     Tee
    0.06
    (page
    0.06
    A
    0.06
    Activations Density 0.152%

    No Known Activations