© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT (Instruct)
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 30169
    Prev
    Next
    INDEX
    Explanations

    disclaimers and recommendations

    np_max-act · gemini-2.0-flash

    phrases or messages related to ethical guidelines about technology use.

    oai_token-act-pair · gpt-4o-miniTriggered by @xinyanhu8

    The neuron activates on policy‐style refusal or disclaimer language listing prohibited or harmful content categories (e.g. “cannot provide answers that are harmful, unethical, racist, sexist, toxic, dangerous, or illegal”).

    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
    .ver
    -0.06
    atre
    -0.06
     monday
    -0.06
     راست
    -0.06
     doit
    -0.06
     ست
    -0.06
     ков
    -0.06
    'o
    -0.06
     دختر
    -0.06
     Schl
    -0.06
    POSITIVE LOGITS
     hundreds
    0.07
    icos
    0.07
     steam
    0.06
    	lines
    0.06
    inci
    0.06
    ặ
    0.06
    //------------------------------------------------------------------------------↵
    0.06
    CTR
    0.06
     //-
    0.06
     symmetric
    0.06
    Activations Density 0.019%

    No Known Activations