© 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. 13015
    Prev
    Next
    INDEX
    Explanations

    markup or syntax elements commonly used in programming or data representation

    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
    388
    +0.15
    0.8%
    435
    +0.11
    0.6%
    375
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    388
    +0.15
    0.03
    203
    +0.11
    0.04
    311
    +0.11
    0.04
    Negative Logits
    ophone
    -1.57
    cule
    -1.42
    aire
    -1.36
    ollah
    -1.36
    cules
    -1.32
    ials
    -1.30
    ycl
    -1.22
     Denmark
    -1.20
    inski
    -1.19
    sc
    -1.18
    POSITIVE LOGITS
     T
    4.19
    ĩ
    3.88
    ľĵ
    3.62
    T
    3.55
    Ĺ
    3.54
    ÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤÃĥÃĤ
    3.49
    ¤
    3.46
    ĭ
    3.46
    č↵
    3.34
    č↵
    3.34
    Activations Density 1.747%

    No Known Activations