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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    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. 88830
    Prev
    Next
    INDEX
    Explanations

    Business development

    np_max-act · gemini-2.0-flash

    The neuron selectively activates on present‐tense action verbs—especially gerunds—describing professional tasks or responsibilities (e.g., planning, executing, managing, capturing, monitoring).

    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
     것은
    -0.07
    eras
    -0.07
    ीछ
    -0.07
     blí
    -0.06
     хочу
    -0.06
    ホ
    -0.06
    hf
    -0.06
     prophet
    -0.06
    RSA
    -0.06
    에
    -0.06
    POSITIVE LOGITS
     Majority
    0.08
    _parm
    0.06
    celik
    0.06
     corner
    0.06
    PASS
    0.06
    -eff
    0.06
     portrayed
    0.06
    176
    0.06
    subseteq
    0.06
     stemmed
    0.06
    Activations Density 0.047%

    No Known Activations