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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 26547
    Prev
    Next
    INDEX
    Explanations

    terms related to multiple sclerosis and its medical context

    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
    446
    +0.13
    0.7%
    275
    +0.13
    0.7%
    414
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    61
    +0.13
    0.08
    448
    +0.13
    0.03
    23
    +0.11
    0.03
    Negative Logits
    [^
    -1.94
    atar
    -1.59
    ressor
    -1.57
    eful
    -1.51
    oclast
    -1.47
     mattered
    -1.46
     \[*
    -1.45
     hypothesis
    -1.45
    struct
    -1.43
     exists
    -1.43
    POSITIVE LOGITS
    ª
    2.60
    ³
    2.56
    ĻĤ
    2.55
    į
    2.52
    Ł
    2.50
    Ĵ
    2.47
    ¯
    2.47
    §
    2.46
    Ĭ
    2.46
    ¼
    2.37
    Activations Density 1.874%

    No Known Activations