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. 85487
    Prev
    Next
    INDEX
    Explanations

    per

    np_max-act · gemini-2.0-flash

    The neuron fires on partial or rare word pieces (subword fragments), i.e. tokens that are not full common words but bits of words created by the tokenizer.

    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
    Nd
    -0.07
    attendance
    -0.06
     Summers
    -0.06
     land
    -0.06
    -out
    -0.06
    results
    -0.06
    ahu
    -0.06
    .writer
    -0.06
     hus
    -0.06
     mah
    -0.06
    POSITIVE LOGITS
    .solution
    0.07
    ('+
    0.07
    _authentication
    0.06
    ující
    0.06
    ości
    0.06
    Continue
    0.06
     MOM
    0.06
    идента
    0.06
    olvimento
    0.06
    .arr
    0.06
    Activations Density 0.015%

    No Known Activations