© 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. 22601
    Prev
    Next
    INDEX
    Explanations

    the conjunction "and" used in various contexts

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    90
    +0.15
    0.8%
    146
    +0.12
    0.6%
    366
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    467
    +0.15
    0.03
    274
    +0.12
    0.05
    208
    +0.11
    0.04
    Negative Logits
     nowhere
    -1.79
     Scot
    -1.67
     Canadians
    -1.55
     pounds
    -1.52
     RSV
    -1.46
     Miss
    -1.43
     inches
    -1.43
     coffee
    -1.42
     reservations
    -1.36
     tall
    -1.34
    POSITIVE LOGITS
    ESULT
    1.81
    probe
    1.78
    SUCCESS
    1.51
    RESULT
    1.48
    args
    1.42
    proc
    1.42
    DLL
    1.41
    dos
    1.39
    control
    1.38
    ource
    1.37
    Activations Density 0.359%

    No Known Activations