© 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. 1-RES-SM
    11. 30091
    Prev
    Next
    INDEX
    Explanations

    the presence of the word "Go" 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/1-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.1.hook_resid_post
    Hook Layer
    1
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Neuron Alignment
    Index
    Value
    % of L₁
    477
    +0.13
    0.7%
    130
    +0.12
    0.6%
    174
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    477
    +0.13
    0.03
    130
    +0.12
    0.03
    399
    +0.11
    0.02
    Negative Logits
    uality
    -1.92
    ships
    -1.86
    ness
    -1.77
    aneous
    -1.61
    ly
    -1.53
    ship
    -1.50
    bury
    -1.49
     times
    -1.45
    GTH
    -1.45
    rough
    -1.44
    POSITIVE LOGITS
    ÅĽci
    1.96
    ppa
    1.82
    ÅĽÄĩ
    1.80
    alt
    1.68
    ppo
    1.66
    RNAs
    1.65
    Äij
    1.62
    etting
    1.57
    eb
    1.54
    epi
    1.50
    Activations Density 0.084%

    No Known Activations