© 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. 129862
    Prev
    Next
    INDEX
    Explanations

    hurt

    np_max-act · gemini-2.0-flash

    phrases related to causing harm or damage.

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

    The neuron consistently flags expressions about causing harm or injury to someone.

    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
     Tham
    -0.07
    stage
    -0.07
    version
    -0.07
     блю
    -0.07
     classical
    -0.06
    اکم
    -0.06
     pageSize
    -0.06
     absolute
    -0.06
    12
    -0.06
     FAIL
    -0.06
    POSITIVE LOGITS
     hurt
    0.11
     Hurt
    0.11
     hurting
    0.10
     hurts
    0.09
    urt
    0.09
    URT
    0.07
     hurd
    0.07
    贴
    0.07
     cut
    0.07
     uh
    0.07
    Activations Density 0.007%

    No Known Activations