© 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. 4-RES-SM
    11. 14139
    Prev
    Next
    INDEX
    Explanations

    patterns related to programming syntax and code structure

    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
    274
    +0.13
    0.7%
    194
    +0.12
    0.6%
    356
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    203
    +0.13
    0.02
    405
    +0.12
    0.02
    497
    +0.11
    0.01
    Negative Logits
     Kirby
    -1.90
     Norfolk
    -1.81
     pirate
    -1.73
     sod
    -1.67
     Jamaica
    -1.66
     waters
    -1.60
     Sri
    -1.60
     reef
    -1.59
     Lynn
    -1.59
     island
    -1.58
    POSITIVE LOGITS
     aut
    3.60
     Aut
    3.45
     Auto
    3.17
    Aut
    3.07
    aut
    2.96
     auto
    2.90
    Auto
    2.84
     AUT
    2.73
    AUT
    2.63
     autophagy
    2.49
    Activations Density 0.662%

    No Known Activations