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. 40568
    Prev
    Next
    INDEX
    Explanations

    could not find a pattern.will

    np_max-act · gemini-2.0-flash

    The neuron is primarily detecting auxiliary and modal verbs—especially “has” and “will.”

    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
    بود
    -0.07
    -track
    -0.07
    /list
    -0.06
    swith
    -0.06
    .Bind
    -0.06
    200
    -0.06
    938
    -0.06
     impartial
    -0.06
     Supplies
    -0.06
     unicorn
    -0.06
    POSITIVE LOGITS
     Sega
    0.07
     sie
    0.06
     کارگرد
    0.06
    formik
    0.06
    =”
    0.06
     nelle
    0.06
     SSR
    0.06
     seni
    0.06
     çalışma
    0.06
     compét
    0.06
    Activations Density 0.172%

    No Known Activations