© 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
    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. 72777
    Prev
    Next
    INDEX
    Explanations

    violence

    np_max-act · gemini-2.0-flash

    The neuron detects legal references to violent offenses—words classifying crimes as “violent felonies,” “crimes of violence,” “aggravated assault,” etc., in sentencing contexts.

    oai_token-act-pair · o4-miniTriggered by @xinyanhu8

    legal terminology related to criminal cases and sentencing.

    oai_token-act-pair · gpt-4o-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
    _ui
    -0.06
     osc
    -0.06
    \Configuration
    -0.06
     Králové
    -0.06
     WF
    -0.06
    参
    -0.05
     abc
    -0.05
    gui
    -0.05
     запис
    -0.05
    编辑
    -0.05
    POSITIVE LOGITS
     Too
    0.07
    ograms
    0.07
    .ResultSet
    0.06
    vironments
    0.06
    hall
    0.06
     Cartoon
    0.06
    ={{↵
    0.06
    ialized
    0.06
     İng
    0.06
     underwent
    0.06
    Activations Density 0.002%

    No Known Activations