© 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. 4-RES-SM
    11. 14660
    Prev
    Next
    INDEX
    Explanations

    Java classes and their methods relevant to data handling and processing

    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/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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    78
    +0.17
    0.9%
    397
    +0.14
    0.8%
    453
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    17
    +0.17
    0.07
    453
    +0.14
    0.08
    484
    +0.12
    0.05
    Negative Logits
    ycin
    -1.88
    manship
    -1.82
    ensis
    -1.64
    liography
    -1.64
     ([@
    -1.61
    hores
    -1.59
    blast
    -1.56
    [\#
    -1.55
    nos
    -1.54
    romycin
    -1.54
    POSITIVE LOGITS
    ¬
    3.77
    µ
    3.72
    ª
    3.57
    ¦
    3.49
    °
    3.48
    ³
    3.47
    ¶
    3.47
    «
    3.46
    ¡
    3.46
    ı
    3.41
    Activations Density 0.763%

    No Known Activations