© 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. 10857
    Prev
    Next
    INDEX
    Explanations

    references to file paths and system-related imports

    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
    156
    +0.20
    1.1%
    142
    +0.14
    0.8%
    94
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    203
    +0.20
    0.03
    142
    +0.14
    0.03
    311
    +0.13
    0.03
    Negative Logits
    ically
    -1.92
    apine
    -1.66
    arius
    -1.65
    inated
    -1.62
    ERY
    -1.61
    ized
    -1.61
    aneously
    -1.55
    naire
    -1.52
    TON
    -1.49
    ITED
    -1.49
    POSITIVE LOGITS
    Ĥ
    4.27
    ĨĴ
    4.17
    ¿
    4.15
    ¶
    4.14
    ¬
    4.10
    ·
    4.07
    Į
    4.04
    IJ
    4.00
    ¼
    3.99
    ij
    3.99
    Activations Density 0.185%

    No Known Activations