© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Olmo-3-1125-32B
    5. 32-RES-BATCHTOPK-131K
    7. 27665
    Prev
    Next
    INDEX
    Explanations

    can/may/will followed by verb or adverb

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    bcywinski/Olmo-3-32B-Base-SAE/saes_allenai_Olmo-3-1125-32B_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
     somehow
    -0.13
     indeed
    -0.12
     surely
    -0.12
     admittedly
    -0.12
     undoubtedly
    -0.11
     outnumber
    -0.11
     obviously
    -0.11
     truly
    -0.11
     ultimately
    -0.10
    xed
    -0.10
    POSITIVE LOGITS
     helpful
    0.20
     help
    0.16
     suffice
    0.15
    help
    0.15
     helps
    0.13
     Helpful
    0.12
     effective
    0.12
     useful
    0.12
     works
    0.11
     helf
    0.11
    Activations Density 0.161%

    No Known Activations