© 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. 91654
    Prev
    Next
    INDEX
    Explanations

    er

    np_max-act · gemini-2.0-flash

    The neuron activates on mentions of anger or aggressive/emotional hostility (e.g. “anger,” “angry,” “aggression”).

    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
     ebook
    -0.07
     follic
    -0.07
    -0.07
    702
    -0.07
     touted
    -0.07
     نو
    -0.06
    experience
    -0.06
     coincide
    -0.06
    44
    -0.06
    15
    -0.06
    POSITIVE LOGITS
     anger
    0.12
     angry
    0.10
     rage
    0.07
     Angry
    0.07
     άλ
    0.07
    ující
    0.07
    _KERNEL
    0.07
    .opacity
    0.06
    PY
    0.06
     geniş
    0.06
    Activations Density 0.007%

    No Known Activations