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. 7-RESID-POST-AA
    6. 5159
    Prev
    Next
    INDEX
    Explanations

    lama

    np_max-act · gemini-2.0-flash

    It detects mentions of the Llama language model name (and its letter-case/variant tokenizations).

    oai_token-act-pair · gpt-5-miniTriggered by @yooniel31
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_7/trainer_1
    Dataset (Dashboard)
    Various
    Features
    131,072
    Data Type
    float32
    Hook Name
    blocks.7.hook_resid_post
    Architecture
    standard
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    _ud
    -0.08
     forge
    -0.08
     deposits
    -0.07
    -reviewed
    -0.07
     survives
    -0.07
    ategorized
    -0.06
    顾
    -0.06
     received
    -0.06
     поверх
    -0.06
     marginalized
    -0.06
    POSITIVE LOGITS
    mare
    0.08
    안마
    0.07
     кожного
    0.06
    čet
    0.06
    _COMM
    0.06
    χώ
    0.06
    _factory
    0.06
     arithmetic
    0.06
    name
    0.06
     Enums
    0.06
    Activations Density 0.001%

    No Known Activations