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. 27380
    Prev
    Next
    INDEX
    Explanations

    references to scientific articles or research citations

    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
    111
    +0.13
    0.7%
    310
    +0.11
    0.6%
    148
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.13
    0.01
    469
    +0.11
    0.00
    339
    +0.11
    0.00
    Negative Logits
     corresponding
    -1.63
     computed
    -1.61
     cross
    -1.55
    izer
    -1.54
     predicted
    -1.53
     percentage
    -1.50
     uptake
    -1.47
     converted
    -1.47
     value
    -1.46
     equivalent
    -1.45
    POSITIVE LOGITS
    »¿
    2.75
    ĻĤ
    2.17
    Īĺ
    2.15
    ¥
    2.03
    ¤
    2.00
    į
    1.96
    Ĩ
    1.95
    ĩ
    1.84
    rums
    1.84
    ¬
    1.81
    Activations Density 0.090%

    No Known Activations