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. 492
    Prev
    Next
    INDEX
    Explanations

    possibility

    np_max-act · gemini-2.0-flash

    The neuron detects speculative or modal auxiliary verbs (e.g., “could,” “may,” “might”) indicating possibility or conjecture.

    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
     isSuccess
    -0.07
    _schedule
    -0.07
    (mesh
    -0.07
     kop
    -0.07
    (ps
    -0.06
    _tls
    -0.06
    tenant
    -0.06
    _mes
    -0.06
     ores
    -0.06
    Len
    -0.06
    POSITIVE LOGITS
     will
    0.11
     may
    0.09
     Will
    0.08
     Could
    0.08
     Can
    0.08
     can
    0.08
    will
    0.07
     would
    0.07
    .Can
    0.07
     could
    0.07
    Activations Density 0.142%

    No Known Activations