© 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. 8679
    Prev
    Next
    INDEX
    Explanations

    the phrase "what" in various contexts, indicating a focus on inquiry or questioning

    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/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
    Neuron Alignment
    Index
    Value
    % of L₁
    307
    +0.16
    0.9%
    320
    +0.13
    0.7%
    77
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    77
    +0.16
    0.07
    390
    +0.13
    0.06
    256
    +0.12
    0.05
    Negative Logits
    İ
    -2.21
    »¿
    -1.85
    µ
    -1.75
    «
    -1.74
    ĻĤ
    -1.73
    OSS
    -1.71
    Ŀ
    -1.67
     whom
    -1.66
    ľ
    -1.64
    mith
    -1.62
    POSITIVE LOGITS
     happens
    2.73
     happened
    2.35
    soever
    2.21
     appears
    2.10
     else
    2.03
     constitutes
    1.98
     follows
    1.92
     we
    1.92
     they
    1.85
     amounts
    1.78
    Activations Density 0.169%

    No Known Activations