© 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. 2-RES-SM
    11. 10175
    Prev
    Next
    INDEX
    Explanations

    prepositions and their associated phrases within the text

    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/2-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.2.hook_resid_post
    Hook Layer
    2
    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
    136
    +0.14
    0.8%
    23
    +0.13
    0.7%
    54
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    440
    +0.14
    0.04
    238
    +0.13
    0.04
    300
    +0.12
    0.04
    Negative Logits
    ī
    -1.77
    writer
    -1.62
     calculator
    -1.57
    chen
    -1.57
    gin
    -1.56
    stown
    -1.52
    mith
    -1.52
     Engineer
    -1.50
    gun
    -1.47
     Racing
    -1.45
    POSITIVE LOGITS
     response
    1.93
    versus
    1.89
     relation
    1.86
    semin
    1.78
    ialize
    1.76
    ferential
    1.73
     order
    1.72
    %).
    1.66
     spite
    1.61
    \%$
    1.60
    Activations Density 0.292%

    No Known Activations