© 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. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 4-RES-SM
    11. 11965
    Prev
    Next
    INDEX
    Explanations

    punctuation marks and their usage in sentences

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/4-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.4.hook_resid_post
    Hook Layer
    4
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    397
    +0.17
    1.0%
    307
    +0.15
    0.9%
    25
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    377
    +0.17
    0.06
    58
    +0.15
    0.07
    72
    +0.14
    0.06
    Negative Logits
    Ļª
    -4.39
    ¼
    -4.24
    ¾
    -4.14
    ĨĴ
    -4.12
    ¶
    -4.07
    ĵ
    -4.04
    ģ
    -3.98
    ¿
    -3.96
    ĥ
    -3.95
    Ĥ¬
    -3.94
    POSITIVE LOGITS
     namely
    2.11
     however
    2.08
     but
    2.06
     коÑĤоÑĢ
    2.04
     my
    1.99
     ÑĩÑĤо
    1.99
     though
    1.93
     although
    1.78
    so
    1.76
    etc
    1.76
    Activations Density 0.395%

    No Known Activations