© 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. 4-RES-SM
    11. 8937
    Prev
    Next
    INDEX
    Explanations

    words and phrases related to cultural identity and representation

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/4-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.4.hook_resid_post
    Hook Layer
    4
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    306
    +0.18
    1.1%
    36
    +0.14
    0.8%
    348
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    203
    +0.18
    0.03
    174
    +0.14
    0.03
    469
    +0.12
    0.04
    Negative Logits
    atory
    -1.98
    ations
    -1.89
    ation
    -1.86
    uous
    -1.76
    ating
    -1.70
    ated
    -1.68
    ized
    -1.58
    uated
    -1.56
    GTH
    -1.55
    uation
    -1.52
    POSITIVE LOGITS
    Ĺ
    3.25
    ¹
    3.04
    Ŀ
    2.86
    ±
    2.82
    ¯
    2.73
    ©
    2.73
    ŀ
    2.70
    Ĥ
    2.67
    Ĵ
    2.66
    ·
    2.62
    Activations Density 0.373%

    No Known Activations