© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 15037
    Prev
    Next
    INDEX
    Explanations

    phrases related to mathematical calculations and problem-solving

    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
    306
    +0.20
    1.2%
    348
    +0.16
    0.9%
    255
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    177
    +0.20
    0.01
    34
    +0.16
    0.01
    501
    +0.12
    0.00
    Negative Logits
    gent
    -1.61
    pace
    -1.40
    idal
    -1.39
    ward
    -1.36
    win
    -1.35
    orer
    -1.34
    nergy
    -1.34
    ctions
    -1.34
    inin
    -1.33
    tti
    -1.30
    POSITIVE LOGITS
     prob
    2.10
    Į
    1.87
    ¤
    1.71
    Ī
    1.63
    Ŀ
    1.54
    Ł
    1.49
    ·
    1.48
    Ļ
    1.46
    ľ
    1.45
    ·¸
    1.44
    Activations Density 0.025%

    No Known Activations