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. 40513
    Prev
    Next
    INDEX
    Explanations

    -old

    np_max-act · gemini-2.0-flash

    The neuron fires on occurrences of the token “censored,” i.e. parts of the text that signal content filtering or refusal.

    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
     discomfort
    -0.07
     شهرد
    -0.07
     loaded
    -0.06
     bury
    -0.06
    kw
    -0.06
    _bal
    -0.06
     gust
    -0.06
     testName
    -0.06
     Sb
    -0.06
     Briggs
    -0.06
    POSITIVE LOGITS
    定
    0.07
    \Base
    0.07
     arbitr
    0.06
    acea
    0.06
    (笑
    0.06
    ίζει
    0.06
     ayr
    0.06
     entsprech
    0.06
     Brett
    0.06
    ญ
    0.06
    Activations Density 0.005%

    No Known Activations