© 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. 2-RES-SM
    11. 14387
    Prev
    Next
    INDEX
    Explanations

    numerical values and structured data formats

    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/2-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.2.hook_resid_post
    Hook Layer
    2
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    240
    +0.15
    0.9%
    479
    +0.15
    0.9%
    243
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    243
    +0.15
    0.16
    125
    +0.15
    0.05
    91
    +0.15
    0.11
    Negative Logits
    ¿½
    -1.69
    ħ
    -1.53
    º
    -1.52
    ¾
    -1.51
    etch
    -1.45
    ncia
    -1.45
     Handbook
    -1.41
    yses
    -1.39
     quo
    -1.37
    hip
    -1.36
    POSITIVE LOGITS
     =>
    1.77
     integers
    1.72
     important
    1.53
     territory
    1.46
     integer
    1.44
     yourself
    1.44
     contiguous
    1.39
     correspond
    1.39
     crucial
    1.37
     biggest
    1.37
    Activations Density 5.630%

    No Known Activations