© 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. MLP Post
    9. 2-MLP-SM
    11. 22713
    Prev
    Next
    INDEX
    Explanations

    technical terms related to inventions or apparatuses

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__mlp-sm_processed/2-mlp-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_mlp_out
    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
    98
    +0.13
    0.7%
    100
    +0.12
    0.7%
    136
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    59
    +0.13
    -0.00
    368
    +0.12
    0.00
    461
    +0.12
    0.01
    Negative Logits
    ª
    -1.65
    ĻĤ
    -1.61
     racism
    -1.58
     advance
    -1.51
     racist
    -1.38
     outstanding
    -1.36
    PRESS
    -1.34
     antif
    -1.33
    ¡
    -1.30
    ļ
    -1.29
    POSITIVE LOGITS
    sole
    1.63
    compare
    1.50
     themselves
    1.46
    lei
    1.45
    assembled
    1.41
    ixel
    1.41
    qui
    1.40
     homolog
    1.36
     ('
    1.35
    laid
    1.35
    Activations Density 0.083%

    No Known Activations