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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Andy Arditi · Finding Misaligned Persona Features in Open-Weight Models
    5. Llama3.1-8B-IT
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 127187
    Prev
    Next
    INDEX
    Explanations

    as

    np_max-act · gemini-2.0-flash

    The neuron detects negative‐judgment words or phrases (e.g. “offensive,” “disrespectful,” “creepy,” “aggressive”) that warn how actions or statements may be perceived.

    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
    /static
    -0.08
     getToken
    -0.07
    ，其中
    -0.07
    CLOCKS
    -0.07
     NSLog
    -0.06
     decoder
    -0.06
    provide
    -0.06
    ')+
    -0.06
    igmoid
    -0.06
     faz
    -0.06
    POSITIVE LOGITS
    idar
    0.07
    検
    0.06
     agony
    0.06
    ダ
    0.06
    ào
    0.06
     yür
    0.06
    .getTarget
    0.06
    断
    0.06
     vigor
    0.06
    finder
    0.06
    Activations Density 0.013%

    No Known Activations