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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-32B
    5. 32-RESID-BATCHTOPK-65K
    7. 20664
    Prev
    Next
    INDEX
    Explanations

    constructive feedback and reinforcement

    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
    èĽĽ
    -0.10
    åĢĴ
    -0.09
    ç¿ĺ
    -0.09
    ewire
    -0.09
    ä½ľæĸĩ
    -0.09
    èĬĭ
    -0.09
    èĩ³æѤ
    -0.09
    å¸Ń
    -0.09
    rons
    -0.08
    æijĨ
    -0.08
    POSITIVE LOGITS
     constructive
    0.15
    Positive
    0.11
    æľīæĦıä¹ī
    0.10
     productive
    0.10
    finger
    0.10
     positive
    0.09
     Positive
    0.09
     destructive
    0.09
    -positive
    0.09
    destruct
    0.09
    Activations Density 0.056%

    No Known Activations