© 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. 26322
    Prev
    Next
    INDEX
    Explanations

    phrases indicating doubt or conflict

    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
    161
    +0.17
    0.9%
    302
    +0.14
    0.8%
    321
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    14
    +0.17
    0.07
    511
    +0.14
    0.04
    247
    +0.12
    0.07
    Negative Logits
    soever
    -2.04
    Īĺ
    -1.87
    gers
    -1.72
    ł
    -1.70
    acing
    -1.65
    uler
    -1.59
    okia
    -1.55
    Ļª
    -1.55
    ĥ
    -1.53
    awn
    -1.52
    POSITIVE LOGITS
     possibility
    2.33
     difference
    1.84
     chance
    1.81
     discrepancy
    1.57
     tendency
    1.56
     temptation
    1.56
     notion
    1.55
     feeling
    1.52
     shortage
    1.51
     need
    1.48
    Activations Density 0.535%

    No Known Activations