© 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. 91636
    Prev
    Next
    INDEX
    Explanations

    Answers/comments

    np_max-act · gemini-2.0-flash

    This neuron activates on content that triggers the model’s refusal policy—detecting disallowed or unethical requests and the refusal phrases (e.g. “cannot,” “recommend,” “illegal,” “assist”) used to decline them.

    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
    JSImport
    -0.06
    	actual
    -0.06
     راست
    -0.06
    -0.06
    .findElement
    -0.06
    -0.06
    linger
    -0.06
    кими
    -0.06
     τι
    -0.06
    獲
    -0.06
    POSITIVE LOGITS
     и
    0.07
    .GetText
    0.07
     rallying
    0.06
     Predictor
    0.06
     PropertyChanged
    0.06
     launch
    0.06
    =None
    0.06
    inter
    0.06
    \C
    0.06
     Txt
    0.06
    Activations Density 0.059%

    No Known Activations