© 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. 91692
    Prev
    Next
    INDEX
    Explanations

    widely

    np_max-act · gemini-2.0-flash

    The neuron strongly activates on the adverb “widely,” i.e. it picks out mentions of broad or extensive use expressed by that word.

    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
     那
    -0.08
     Cham
    -0.08
     substituted
    -0.07
     commemorate
    -0.07
    üme
    -0.06
    。那
    -0.06
    ([[
    -0.06
     imkan
    -0.06
    兩
    -0.06
     малень
    -0.06
    POSITIVE LOGITS
     widely
    0.13
     ActionListener
    0.07
    IPv
    0.07
     widespread
    0.07
    ERSION
    0.06
    _usage
    0.06
     найбіль
    0.06
    _expected
    0.06
     voting
    0.06
     insider
    0.06
    Activations Density 0.012%

    No Known Activations