© 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. 14279
    Prev
    Next
    INDEX
    Explanations

    dates and timestamps in the text

    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
    111
    +0.12
    0.7%
    335
    +0.12
    0.7%
    115
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    131
    +0.12
    0.05
    425
    +0.12
    0.05
    69
    +0.11
    0.05
    Negative Logits
    ´
    -1.60
     sino
    -1.55
    ī
    -1.53
     digitally
    -1.42
     ever
    -1.42
     drill
    -1.41
     nuclei
    -1.39
     leverage
    -1.37
    Īĺ
    -1.36
     remotely
    -1.35
    POSITIVE LOGITS
    versation
    1.58
    vous
    1.53
    comments
    1.47
    eration
    1.45
    IBILITY
    1.43
    ête
    1.42
     respects
    1.42
    ailability
    1.41
    noon
    1.41
    udeau
    1.41
    Activations Density 0.180%

    No Known Activations