© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 2-RES-SM
    11. 26352
    Prev
    Next
    INDEX
    Explanations

    programming-related terminology and concepts

    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/2-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.2.hook_resid_post
    Hook Layer
    2
    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
    419
    +0.32
    1.9%
    186
    +0.21
    1.2%
    198
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    419
    +0.32
    0.14
    104
    +0.21
    0.08
    21
    +0.14
    0.07
    Negative Logits
    noreply
    -1.59
    axel
    -1.57
    ienna
    -1.45
    llo
    -1.44
    olis
    -1.41
     interest
    -1.37
    letons
    -1.36
    Ļª
    -1.33
    ales
    -1.33
    ents
    -1.29
    POSITIVE LOGITS
    gue
    1.72
    -[
    1.49
    acious
    1.42
    avy
    1.42
    ensch
    1.35
    unov
    1.33
    âĢ¦]
    1.32
    idity
    1.32
    uparrow
    1.31
     eats
    1.31
    Activations Density 3.338%

    No Known Activations