© 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. 1-RES-SM
    11. 24046
    Prev
    Next
    INDEX
    Explanations

    references to software development and related concepts

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    365
    +0.14
    0.8%
    244
    +0.12
    0.7%
    208
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    385
    +0.14
    0.01
    208
    +0.12
    0.02
    453
    +0.11
    0.01
    Negative Logits
     different
    -1.64
    encies
    -1.62
     happier
    -1.61
     remember
    -1.52
     shame
    -1.52
     weighted
    -1.52
     half
    -1.50
     loneliness
    -1.49
    ¯
    -1.47
    most
    -1.46
    POSITIVE LOGITS
    arium
    1.98
    fiddle
    1.70
    aceut
    1.70
    IFIC
    1.63
    ifacts
    1.59
    cript
    1.57
    works
    1.57
     Engineer
    1.56
     Architect
    1.55
     Sources
    1.55
    Activations Density 0.064%

    No Known Activations