© 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. 8565
    Prev
    Next
    INDEX
    Explanations

    references to judicial rulings and legal cases

    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
    +0.25
    1.4%
    48
    +0.13
    0.7%
    168
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    168
    +0.25
    0.09
    10
    +0.13
    0.07
    410
    +0.13
    0.06
    Negative Logits
     optimization
    -1.62
     targeting
    -1.58
     flu
    -1.49
     coding
    -1.44
     batches
    -1.42
     translated
    -1.42
    launchpad
    -1.42
     luck
    -1.38
     segmentation
    -1.38
    üller
    -1.37
    POSITIVE LOGITS
    »¿
    2.81
    Ļª
    2.68
    ĸ
    2.44
    ¬
    2.38
    ķ
    2.34
    ²
    2.30
    ĩ
    2.29
    ĨĴ
    2.25
    Ģ
    2.21
    Īĺ
    2.16
    Activations Density 0.560%

    No Known Activations