© 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. 22627
    Prev
    Next
    INDEX
    Explanations

    references to categories or divisions within a structure

    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
    121
    +0.16
    0.9%
    465
    +0.14
    0.8%
    428
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    465
    +0.16
    0.04
    14
    +0.14
    0.03
    121
    +0.11
    0.02
    Negative Logits
     perturb
    -1.59
     typ
    -1.58
     typedef
    -1.55
    ational
    -1.40
     numerically
    -1.40
    align
    -1.39
     %{
    -1.37
     promul
    -1.36
    points
    -1.34
    binom
    -1.33
    POSITIVE LOGITS
    wagen
    2.12
    bourne
    1.78
    aku
    1.64
    gia
    1.62
    apat
    1.56
    shot
    1.55
     shots
    1.54
    enta
    1.51
     shooter
    1.42
    ships
    1.41
    Activations Density 0.192%

    No Known Activations