© 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. 26941
    Prev
    Next
    INDEX
    Explanations

    forms of the verb "to be" in various tenses

    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
    159
    +0.13
    0.7%
    324
    +0.12
    0.7%
    327
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    324
    +0.13
    0.04
    327
    +0.12
    0.04
    88
    +0.11
    0.02
    Negative Logits
    ties
    -1.49
     lengths
    -1.49
     lacking
    -1.47
    essential
    -1.44
     isolated
    -1.43
     factors
    -1.39
     tops
    -1.38
     indicated
    -1.37
     degrees
    -1.37
    zzo
    -1.36
    POSITIVE LOGITS
    ĨĴ
    3.05
    ¿½
    2.90
    ł
    2.87
    ½
    2.85
    Ķ
    2.75
    ¢
    2.70
    ↵↵
    2.69
    2.69
    ↵↵
    2.69
    ↵↵
    2.69
    Activations Density 0.377%

    No Known Activations