© 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. 15236
    Prev
    Next
    INDEX
    Explanations

    terms and phrases related to measurements and evaluations of efficiency

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    156
    +0.17
    1.0%
    78
    +0.14
    0.8%
    63
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    188
    +0.17
    0.05
    11
    +0.14
    0.08
    85
    +0.13
    0.06
    Negative Logits
     quo
    -1.96
    aders
    -1.68
    sein
    -1.61
     vain
    -1.59
    him
    -1.54
    voir
    -1.53
    oons
    -1.48
    ader
    -1.47
    isenberg
    -1.46
    hin
    -1.45
    POSITIVE LOGITS
     size
    3.41
     number
    3.17
     length
    3.05
     percentage
    3.03
     diameter
    3.03
     amount
    3.03
     rate
    2.93
    ļ
    2.89
     sizes
    2.86
    §
    2.85
    Activations Density 2.388%

    No Known Activations