Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    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. 127549
    Prev
    Next
    INDEX
    Explanations

    login

    np_max-act · gemini-2.0-flash

    Explanation of neuron 4 behavior: the main thing this neuron does is find numerical tokens (digits or numbers).

    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
     trainers
    -0.07
     Withdraw
    -0.07
    demo
    -0.06
     стара
    -0.06
     discriminator
    -0.06
    igua
    -0.06
    ath
    -0.06
     bells
    -0.06
     bien
    -0.06
    較
    -0.06
    POSITIVE LOGITS
    =False
    0.08
    GROUND
    0.07
    Bron
    0.07
     frightening
    0.07
     FRONT
    0.06
    .Identifier
    0.06
    aped
    0.06
     Usually
    0.06
    olean
    0.06
    atisfied
    0.06
    Activations Density 0.015%

    No Known Activations