© 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
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 55481
    Prev
    Next
    INDEX
    Explanations

    oin

    np_max-act · gemini-2.0-flash

    The neuron is a “proper-noun” detector—it activates on capitalized tokens (names, titles, headings, etc.), especially when they’re repeated or occur outside ordinary sentence starts.

    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
     £
    -0.07
    (node
    -0.06
    Marshal
    -0.06
    ALLOW
    -0.06
     skiing
    -0.06
    YSTEM
    -0.06
     went
    -0.06
    apyrus
    -0.06
    historic
    -0.06
    .NO
    -0.06
    POSITIVE LOGITS
    /column
    0.08
    #ga
    0.07
     للم
    0.07
    τικές
    0.06
     ttk
    0.06
     haute
    0.06
    _Callback
    0.06
    .toDouble
    0.06
     oldu
    0.06
     magna
    0.06
    Activations Density 0.055%

    No Known Activations