© 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. 26257
    Prev
    Next
    INDEX
    Explanations

    phrases emphasizing comparisons and contrasts

    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
    291
    +0.18
    1.0%
    478
    +0.16
    1.0%
    56
    +0.16
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    291
    +0.18
    0.26
    10
    +0.16
    0.23
    98
    +0.16
    0.03
    Negative Logits
    "}](#
    -1.67
    abling
    -1.52
    Ä
    -1.50
    gue
    -1.48
    ward
    -1.48
     Boltzmann
    -1.43
    uchi
    -1.40
    gment
    -1.39
    dorff
    -1.38
     among
    -1.38
    POSITIVE LOGITS
     likes
    2.70
    ĻĤ
    2.35
     same
    2.33
     whole
    2.19
    rapeut
    2.13
     slightest
    2.11
     latter
    2.11
     guy
    2.10
     guys
    2.07
    ses
    2.04
    Activations Density 2.842%

    No Known Activations