Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    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. 12821
    Prev
    Next
    INDEX
    Explanations

    topic/subject

    np_max-act · gemini-2.0-flash

    The neuron is triggered by the Russian phrase “на тему,” i.e. requests specifying “on the topic” of something.

    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
     blush
    -0.07
    script
    -0.07
    _unit
    -0.07
    dir
    -0.06
     personel
    -0.06
    BU
    -0.06
     bulunuyor
    -0.06
    Loop
    -0.06
     출연
    -0.06
    HOUSE
    -0.06
    POSITIVE LOGITS
    ุงเทพ
    0.07
     quaint
    0.07
    assic
    0.06
     tutorial
    0.06
    .connector
    0.06
     még
    0.06
     sanitation
    0.06
    .replaceAll
    0.06
    RetVal
    0.06
    _ip
    0.06
    Activations Density 0.023%

    No Known Activations