© 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. 39013
    Prev
    Next
    INDEX
    Explanations

    poor followed by negative outcomes or states

    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
    ç»ıè´¸
    -0.08
    .VAL
    -0.08
    rightness
    -0.08
    æĽ´å¤ļ
    -0.08
    ç¾İå¾·
    -0.08
    hed
    -0.08
    è©
    -0.07
    饬
    -0.07
    adel
    -0.07
     Tradable
    -0.07
    POSITIVE LOGITS
    -quality
    0.13
    é²ł
    0.12
     excuse
    0.10
     quality
    0.10
    æĭĽå¾ħ
    0.10
    canf
    0.10
    èĩªèIJ¥
    0.09
    /no
    0.09
    ½æķ°
    0.09
    åĩłæŃ¥
    0.09
    Activations Density 0.029%

    No Known Activations