© 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. 8762
    Prev
    Next
    INDEX
    Explanations

    patterns in URLs and web domain structures

    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
    78
    +0.15
    0.9%
    172
    +0.13
    0.8%
    40
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    255
    +0.15
    0.00
    172
    +0.13
    0.03
    78
    +0.10
    0.00
    Negative Logits
     neither
    -1.40
     theirs
    -1.37
     lesser
    -1.35
    +$
    -1.34
    ize
    -1.30
    ilate
    -1.30
    ulate
    -1.29
    igned
    -1.26
    nia
    -1.22
     nothing
    -1.21
    POSITIVE LOGITS
    Ģ
    2.74
    Ķ
    2.57
    ĨĴ
    2.50
    İ
    2.49
    ĥ
    2.48
    IJ
    2.48
    ģ
    2.48
    ĥ½
    2.45
    Ĥ
    2.44
    ³
    2.43
    Activations Density 0.136%

    No Known Activations