© 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. 32333
    Prev
    Next
    INDEX
    Explanations

    information related to technical or programming concepts and elements

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    17
    +0.20
    1.1%
    21
    +0.13
    0.7%
    407
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    407
    +0.20
    0.17
    247
    +0.13
    0.19
    358
    +0.13
    0.20
    Negative Logits
    iations
    -1.66
    comings
    -1.46
    Acknowled
    -1.46
    Ļª
    -1.42
    raints
    -1.40
    igation
    -1.39
     insights
    -1.38
    ArgumentException
    -1.38
    anguage
    -1.36
    Methods
    -1.36
    POSITIVE LOGITS
     _____
    1.49
     grocer
    1.45
     same
    1.38
    cher
    1.32
    lickr
    1.31
     £
    1.29
     sold
    1.24
    amous
    1.22
     gallon
    1.21
     './
    1.20
    Activations Density 7.047%

    No Known Activations