© 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
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 90783
    Prev
    Next
    INDEX
    Explanations

    articles "a" and "an"

    np_max-act · gemini-2.0-flash

    The neuron activates on evaluative quality descriptors—positive adjectives and adverbs (e.g. “quick,” “efficient,” “healthy,” “positive,” “full-time”) used in formal or managerial language.

    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
    times
    -0.07
    Passed
    -0.07
    preci
    -0.06
    -moving
    -0.06
    Workflow
    -0.06
    -hit
    -0.06
    hcp
    -0.06
    Road
    -0.06
    _f
    -0.06
    -fold
    -0.06
    POSITIVE LOGITS
    _SEG
    0.06
     UINavigationController
    0.06
    時間
    0.06
     uploader
    0.06
     ارتباط
    0.06
     ก
    0.06
     compatibility
    0.06
     máte
    0.06
    actics
    0.06
     첨부파일
    0.06
    Activations Density 0.013%

    No Known Activations