© 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. 1-RES-SM
    11. 24597
    Prev
    Next
    INDEX
    Explanations

    conjunctions and their frequency in connecting phrases

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    23
    +0.24
    1.4%
    156
    +0.22
    1.3%
    494
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    494
    +0.24
    0.11
    111
    +0.22
    0.13
    61
    +0.11
    0.10
    Negative Logits
    ipel
    -1.73
     alone
    -1.59
     yourself
    -1.55
    place
    -1.50
     info
    -1.49
    äsident
    -1.45
    ”.
    -1.43
    criptor
    -1.43
     pad
    -1.42
    pad
    -1.41
    POSITIVE LOGITS
    ¯
    3.71
    ħ
    3.19
    ª
    3.04
    ¨
    2.88
    Ļª
    2.80
    ¿½
    2.80
    ³
    2.79
    IJ
    2.75
    ¿
    2.71
    ´
    2.65
    Activations Density 1.512%

    No Known Activations