Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    3. Llama3.1-8B-IT (Instruct)
    4. Resid Post - 131k
    5. 11-RESID-POST-AA
    6. 85868
    Prev
    Next
    INDEX
    Explanations

    Quiet, gentle actions

    np_max-act · gemini-2.0-flash

    This neuron specializes in detecting adverbs of manner (words ending in “-ly,” especially “softly,” “quietly,” 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
     Welt
    -0.06
     больш
    -0.06
    felt
    -0.06
     딸
    -0.06
     nhỏ
    -0.06
                                                                    
    -0.06
     Fro
    -0.06
     họa
    -0.06
     květ
    -0.06
     GUIStyle
    -0.06
    POSITIVE LOGITS
    ути
    0.07
    anguage
    0.07
    ọ
    0.07
     solidarity
    0.07
    ransition
    0.06
     anniversary
    0.06
    likelihood
    0.06
     gender
    0.06
     hect
    0.06
    ữ
    0.06
    Activations Density 0.011%

    No Known Activations