© 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. 32304
    Prev
    Next
    INDEX
    Explanations

    actions related to accessibility and usage in various contexts

    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
    500
    +0.17
    1.0%
    358
    +0.13
    0.7%
    274
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    500
    +0.17
    0.13
    358
    +0.13
    0.14
    148
    +0.13
    0.03
    Negative Logits
    iator
    -2.06
    §
    -1.81
    hedral
    -1.50
    heit
    -1.46
    iors
    -1.46
    ŀ
    -1.43
    ¹
    -1.37
    bach
    -1.36
    "}](#
    -1.36
    chter
    -1.34
    POSITIVE LOGITS
     donations
    1.46
     behalf
    1.45
     possession
    1.42
     Nigerian
    1.36
    Thursday
    1.36
     purposes
    1.35
    ês
    1.35
     affairs
    1.35
     coff
    1.31
     unpaid
    1.30
    Activations Density 1.563%

    No Known Activations