© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 9640
    Prev
    Next
    INDEX
    Explanations

    classes and methods related to programming and data structures

    oai_token-act-pair · gpt-4o-miniTriggered 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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    156
    +0.43
    2.8%
    369
    +0.18
    1.1%
    240
    +0.16
    1.0%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    369
    +0.43
    0.05
    379
    +0.18
    0.10
    240
    +0.16
    0.00
    Negative Logits
     theorem
    -1.38
     circum
    -1.36
    sing
    -1.34
     rejection
    -1.33
    know
    -1.29
     majority
    -1.29
    EX
    -1.27
     contrast
    -1.25
    laid
    -1.24
    su
    -1.24
    POSITIVE LOGITS
    ĨĴ
    2.96
    ¢
    2.90
    Ļª
    2.69
    ļ
    2.67
    ĭ
    2.67
    ³
    2.67
    ¨
    2.66
    Īĺ
    2.65
    ¦
    2.65
      ↵
    2.62
    Activations Density 1.197%

    No Known Activations