© 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. 8896
    Prev
    Next
    INDEX
    Explanations

    Suspicion

    np_max-act · gemini-2.0-flash

    This neuron detects Portuguese words related to mistrust or distrust (e.g. “desconfiança,” “desconfiado,” “desconfiante”).

    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
     свят
    -0.06
    徽
    -0.06
    -License
    -0.06
    creates
    -0.06
     Newcastle
    -0.06
    gp
    -0.06
    _Invoke
    -0.06
    能
    -0.06
    Unhandled
    -0.06
    Proveedor
    -0.06
    POSITIVE LOGITS
     suspicions
    0.12
     suspicion
    0.10
     paranoia
    0.09
     paranoid
    0.09
     wary
    0.09
     suspicious
    0.09
     distrust
    0.08
     aggressively
    0.07
    ίδα
    0.07
     tard
    0.07
    Activations Density 0.013%

    No Known Activations