Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. MLP Post
    5. 0-MLP-SM
    6. 27633
    Prev
    Next
    INDEX
    Explanations

    instances where the word "explain" is used

    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/0-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.0.hook_mlp_out
    Hook Layer
    0
    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
    376
    +0.21
    1.2%
    115
    +0.15
    0.9%
    372
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    167
    +0.21
    0.02
    372
    +0.15
    0.02
    345
    +0.12
    0.02
    Negative Logits
    ubicin
    -1.75
     brow
    -1.75
    .]{}
    -1.72
    \]]{}
    -1.51
    \])]{}
    -1.47
    )];
    -1.46
    )]{}
    -1.46
    s
    -1.43
    ]{}.
    -1.43
    Bg
    -1.40
    POSITIVE LOGITS
     why
    1.89
     how
    1.55
    ķ
    1.54
    tera
    1.52
    ably
    1.51
     famine
    1.49
    ĸ
    1.44
    Ł
    1.43
    partum
    1.41
     error
    1.35
    Activations Density 1.150%

    No Known Activations