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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 38878
    Prev
    Next
    INDEX
    Explanations

    não

    np_max-act · gemini-2.0-flash

    The neuron primarily detects the Italian negation particle “non” (and its variant “No”) appearing in the text.

    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
    atts
    -0.08
    osaurs
    -0.07
    .Drawing
    -0.07
    ovable
    -0.07
    inces
    -0.07
     Gat
    -0.07
    word
    -0.07
     Gur
    -0.07
    vit
    -0.06
    ukes
    -0.06
    POSITIVE LOGITS
     not
    0.10
    't
    0.09
     não
    0.08
    ’t
    0.08
     не
    0.08
     không
    0.07
     NOT
    0.07
     أن
    0.07
     δεν
    0.07
     NEVER
    0.07
    Activations Density 0.040%

    No Known Activations