© 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. 23976
    Prev
    Next
    INDEX
    Explanations

    references to various powders, specifically in a context related to pharmaceuticals or dietary supplements

    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
    156
    +0.20
    1.2%
    115
    +0.13
    0.7%
    164
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    164
    +0.20
    0.02
    115
    +0.13
    0.01
    392
    +0.13
    0.02
    Negative Logits
    ķ
    -3.46
    Ń
    -3.37
    ĵ
    -3.27
    ĭ
    -3.22
    ĺ
    -3.22
    Ŀ
    -3.20
    ±
    -3.12
    £
    -3.11
    į
    -2.98
    Ĭ
    -2.98
    POSITIVE LOGITS
    ipe
    1.74
    less
    1.73
    soft
    1.69
    itious
    1.68
    brush
    1.68
    ings
    1.66
    illin
    1.65
    ball
    1.65
    mouth
    1.61
    Amount
    1.59
    Activations Density 0.132%

    No Known Activations