© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-32B
    5. 32-RESID-BATCHTOPK-65K
    7. 32570
    Prev
    Next
    INDEX
    Explanations

    literal vs figurative language

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    adamkarvonen/qwen3-32b-saes/saes_Qwen_Qwen3-32B_batch_top_k/resid_post_layer_32
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    éĢŁ
    -0.09
    oten
    -0.09
    åį·
    -0.09
    çªĥ
    -0.08
     fac
    -0.08
    sembl
    -0.08
     Perfect
    -0.08
    åĹľ
    -0.08
    è¡°
    -0.08
    TexCoord
    -0.08
    POSITIVE LOGITS
     metaphor
    0.22
    æ¯Ķåĸ»
    0.20
     literal
    0.18
    literal
    0.17
    åĸ»
    0.16
     literally
    0.15
    å¯ĵæĦı
    0.15
     figur
    0.15
     metaph
    0.14
     Literal
    0.14
    Activations Density 0.051%

    No Known Activations