© 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. 32410
    Prev
    Next
    INDEX
    Explanations

    references to legal cases or court proceedings

    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
    174
    +0.14
    0.8%
    102
    +0.11
    0.6%
    87
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    174
    +0.14
    0.04
    450
    +0.11
    0.04
    319
    +0.10
    0.03
    Negative Logits
     ones
    -1.60
     headlines
    -1.52
     scenes
    -1.50
    obox
    -1.47
    charts
    -1.44
    CNN
    -1.42
     UK
    -1.42
     hasht
    -1.41
    esium
    -1.40
     percentages
    -1.39
    POSITIVE LOGITS
    ´
    2.60
    ¸
    2.44
    ©
    2.35
    ·
    2.32
    ķ
    2.18
    Č
    2.11
    Ĥ
    2.11
    ŀ
    2.11
    °
    2.07
    ¦
    2.05
    Activations Density 0.148%

    No Known Activations