© 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. 23849
    Prev
    Next
    INDEX
    Explanations

    references to analog or analogue concepts 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
    156
    +0.20
    1.1%
    443
    +0.13
    0.7%
    95
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    48
    +0.20
    0.01
    417
    +0.13
    0.01
    443
    +0.12
    0.01
    Negative Logits
     viv
    -1.47
    «
    -1.46
    ĺ
    -1.42
    áĢIJ
    -1.42
     Europa
    -1.39
    áĢ¯
    -1.35
    áĢ¬
    -1.34
     lesser
    -1.29
     cual
    -1.25
    ="../../../../../../
    -1.24
    POSITIVE LOGITS
    ously
    1.79
    izing
    1.73
    ronic
    1.69
    istic
    1.66
    etics
    1.60
    ging
    1.56
    ently
    1.55
    ious
    1.54
    odend
    1.51
    otic
    1.45
    Activations Density 0.090%

    No Known Activations