© 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. 25801
    Prev
    Next
    INDEX
    Explanations

    the presence of the word "ular" in various forms and 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
    98
    +0.14
    0.8%
    492
    +0.13
    0.8%
    59
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.14
    0.02
    452
    +0.13
    0.02
    488
    +0.11
    0.01
    Negative Logits
     awake
    -2.09
     gentle
    -1.62
     arresting
    -1.59
     uh
    -1.59
     wait
    -1.57
     plaintiffs
    -1.56
     rob
    -1.55
     dry
    -1.54
     practical
    -1.50
     civilized
    -1.48
    POSITIVE LOGITS
    µ
    3.86
    º
    3.66
    ¥
    3.62
    Ł
    3.62
    ¿
    3.55
    £
    3.51
    ¢
    3.46
    ´
    3.45
    Ĵ
    3.45
    Ģ
    3.33
    Activations Density 0.068%

    No Known Activations