© 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. 73673
    Prev
    Next
    INDEX
    Explanations

    role-playing game actions

    np_max-act · gemini-2.0-flash

    This neuron selectively activates on words and phrases related to proposing or referring to taking actions (e.g. “actions,” “take,” “what to do next”).

    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
     horrible
    -0.07
    875
    -0.07
     vectors
    -0.07
    benchmark
    -0.07
     infamous
    -0.07
     object
    -0.06
    !("
    -0.06
    염
    -0.06
     ucfirst
    -0.06
     crime
    -0.06
    POSITIVE LOGITS
     cigarette
    0.06
    ेब
    0.06
     імп
    0.06
    排
    0.06
     виб
    0.06
    ільки
    0.06
     bike
    0.06
     fig
    0.06
     dissolved
    0.06
     glitter
    0.05
    Activations Density 0.008%

    No Known Activations