© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 13123
    Prev
    Next
    INDEX
    Explanations

    proper nouns and words associated with various people

    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
    156
    +0.24
    1.4%
    76
    +0.13
    0.7%
    477
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.24
    0.18
    156
    +0.13
    0.15
    71
    +0.12
    0.13
    Negative Logits
    ifically
    -1.76
    ited
    -1.75
    iful
    -1.66
    heer
    -1.66
    its
    -1.64
    acity
    -1.60
    ify
    -1.53
    ¬
    -1.52
    iels
    -1.48
    ¥
    -1.47
    POSITIVE LOGITS
    sis
    2.60
    e
    2.56
    nn
    2.50
    mpi
    2.48
    s
    2.46
    tis
    2.33
    lation
    2.18
    lene
    2.15
    tics
    2.13
    gia
    2.06
    Activations Density 0.732%

    No Known Activations