© 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. 29958
    Prev
    Next
    INDEX
    Explanations

    punctuation

    np_max-act · gemini-2.0-flash

    The neuron fires almost exclusively on the special chat‐format delimiters (e.g. <|eot_id|>, <|end_header_id|>, blank lines) that mark the boundaries between user/assistant turns.

    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
    ADIO
    -0.06
    omer
    -0.06
     Champions
    -0.06
     campos
    -0.06
    (so
    -0.06
    างว
    -0.06
     competitive
    -0.06
     witness
    -0.06
     момент
    -0.05
     decade
    -0.05
    POSITIVE LOGITS
     NYPD
    0.07
     če
    0.07
    _refptr
    0.07
     похож
    0.06
     çalışmalar
    0.06
    ('/:
    0.06
     включа
    0.06
    ユ
    0.06
    textAlign
    0.06
     tt
    0.06
    Activations Density 0.030%

    No Known Activations