© 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 EvalsExportsinterp-engineNEWAPI 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. 23883
    Prev
    Next
    INDEX
    Explanations

    the word "from" in various contexts, indicating it may be detecting relationships or origins

    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
    358
    +0.19
    1.0%
    238
    +0.15
    0.8%
    505
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    358
    +0.19
    0.06
    238
    +0.15
    0.03
    266
    +0.11
    0.00
    Negative Logits
     substrates
    -1.61
    Âł
    -1.61
    amental
    -1.54
     officers
    -1.50
     Nos
    -1.48
    hua
    -1.46
     [@
    -1.45
     $[\
    -1.44
    HA
    -1.43
    á»ģ
    -1.43
    POSITIVE LOGITS
    !'
    1.69
     earning
    1.56
    orrow
    1.55
    '">
    1.51
    Ĵ
    1.46
     lining
    1.46
     fame
    1.45
    !’
    1.44
     '</
    1.43
    ukes
    1.40
    Activations Density 0.932%

    No Known Activations