© 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. 1-RES-SM
    11. 26307
    Prev
    Next
    INDEX
    Explanations

    sequences indicating numeric representations, particularly focusing on the digits of numbers in various contexts

    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
    350
    +0.16
    0.9%
    114
    +0.12
    0.7%
    186
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    350
    +0.16
    0.02
    259
    +0.12
    0.02
    114
    +0.11
    0.02
    Negative Logits
    ODY
    -1.63
    >(
    -1.50
    urane
    -1.49
    elfare
    -1.42
    ven
    -1.40
    áŁ
    -1.38
     benefit
    -1.38
    odor
    -1.37
    ogn
    -1.36
    UD
    -1.34
    POSITIVE LOGITS
    atories
    1.54
    arts
    1.52
    ale
    1.50
     arc
    1.46
     Sin
    1.45
    break
    1.41
    kte
    1.33
    ager
    1.32
     candles
    1.29
     bursts
    1.28
    Activations Density 0.065%

    No Known Activations