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. 130754
    Prev
    Next
    INDEX
    Explanations

    medical research

    np_max-act · gemini-2.0-flash

    The neuron activates on tokens representing numeric quantities or counts (e.g., “13,” “four,” “six,” “eight”).

    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
    akan
    -0.06
    dept
    -0.06
     suggestive
    -0.06
     <?
    -0.06
     saldırı
    -0.06
    ičky
    -0.06
     Premiership
    -0.06
     EX
    -0.06
     yards
    -0.06
    ellation
    -0.06
    POSITIVE LOGITS
    еного
    0.08
    ktion
    0.07
    _internal
    0.07
    memset
    0.06
    leşik
    0.06
    ,)
    0.06
     Proto
    0.06
    amentos
    0.06
    _argv
    0.06
    ogonal
    0.06
    Activations Density 0.008%

    No Known Activations