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

    Neuronpedia

    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. MLP Post
    9. 2-MLP-SM
    11. 25813
    Prev
    Next
    INDEX
    Explanations

    code structure and syntax patterns

    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
    156
    +0.21
    1.2%
    387
    +0.14
    0.8%
    408
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    208
    +0.21
    0.05
    385
    +0.14
    0.05
    115
    +0.11
    0.05
    Negative Logits
    àµį
    -1.68
    seud
    -1.57
    à±į
    -1.50
    APTER
    -1.49
    flix
    -1.47
    rated
    -1.47
    àµ
    -1.43
    iban
    -1.42
    blic
    -1.41
    eer
    -1.39
    POSITIVE LOGITS
     (_
    1.73
    --"
    1.47
     convenience
    1.47
    akers
    1.44
     law
    1.34
     now
    1.33
    '>
    1.32
    stand
    1.32
    my
    1.30
     _.
    1.30
    Activations Density 0.366%

    No Known Activations