© 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. 13514
    Prev
    Next
    INDEX
    Explanations

    Generic English text

    np_max-act · gemini-2.0-flash

    This neuron detects mentions of reactions or unexpected side-effects (words like “react,” “reaction,” “unexpected,” “unpredictable,” “effect,” etc.).

    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
    .created
    -0.07
     发
    -0.07
     payable
    -0.07
     رئیس
    -0.06
    [String
    -0.06
     LocalDate
    -0.06
    ("//*[@
    -0.06
    spec
    -0.06
    �
    -0.06
    handlers
    -0.06
    POSITIVE LOGITS
    furt
    0.07
     θε
    0.07
     白
    0.06
    	ad
    0.06
     dialogRef
    0.06
     atual
    0.06
     tracker
    0.06
     kriz
    0.06
     saves
    0.06
    čast
    0.06
    Activations Density 0.248%

    No Known Activations