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. 40455
    Prev
    Next
    INDEX
    Explanations

    math problems

    np_max-act · gemini-2.0-flash

    The neuron fires on the standalone numeric tokens that represent the computed answer (e.g. the greatest/least common divisor or multiple) in math‐question outputs.

    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
     Gentle
    -0.07
     адже
    -0.07
     sneakers
    -0.06
    byname
    -0.06
    -yellow
    -0.06
    能力
    -0.06
    'n
    -0.06
    \Core
    -0.06
    licated
    -0.06
    كييف
    -0.06
    POSITIVE LOGITS
     kinase
    0.07
    significant
    0.07
     conna
    0.07
     Padres
    0.07
    ึง
    0.06
    vincial
    0.06
    ира
    0.06
    ixed
    0.06
     slash
    0.06
     organ
    0.06
    Activations Density 0.045%

    No Known Activations