© 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
    7. Resid Post - 131k
    9. 11-RESID-POST-AA
    11. 126157
    Prev
    Next
    INDEX
    Explanations

    earthquake magnitude

    np_max-act · gemini-2.0-flash

    Explanation of neuron 4 behavior: the main thing this neuron does is find numeric tokens (especially decimal or multi‐digit numbers such as earthquake magnitudes or dates).

    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
    _transient
    -0.07
    Đ
    -0.06
    brick
    -0.06
     poll
    -0.06
    .oper
    -0.06
    辞
    -0.06
    _Post
    -0.06
     dob
    -0.06
    urate
    -0.06
    عت
    -0.06
    POSITIVE LOGITS
    777
    0.07
    umble
    0.06
    JO
    0.06
     вико
    0.06
    uspended
    0.06
    _inner
    0.06
     Merge
    0.06
    vents
    0.06
    AttributedString
    0.06
     testified
    0.06
    Activations Density 0.003%

    No Known Activations