© 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 EvalsExportsAPI 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. 12601
    Prev
    Next
    INDEX
    Explanations

    instances of the term "deep," often in the context of depth or significance

    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
    198
    +0.15
    0.8%
    350
    +0.13
    0.7%
    429
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    198
    +0.15
    0.04
    429
    +0.13
    0.03
    119
    +0.13
    0.02
    Negative Logits
    ously
    -2.05
    ately
    -1.74
    ably
    -1.72
    nez
    -1.70
    ulin
    -1.70
    ulous
    -1.58
    ureus
    -1.56
    ically
    -1.55
    ricular
    -1.55
    usement
    -1.54
    POSITIVE LOGITS
    Ļ
    2.79
    ľĵ
    2.67
    ening
    2.63
    ĻĤ
    2.63
    ®
    2.60
    Ļª
    2.55
    ĸ´
    2.54
    ĨĴ
    2.51
    ¾
    2.48
    ļ
    2.45
    Activations Density 0.160%

    No Known Activations