© 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
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 1-RES-SM
    6. 26644
    Prev
    Next
    INDEX
    Explanations

    punctuation and formatting elements in written text

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/1-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.1.hook_resid_post
    Hook Layer
    1
    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
    156
    +0.29
    1.7%
    111
    +0.21
    1.2%
    17
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.29
    0.42
    271
    +0.21
    0.18
    156
    +0.15
    0.23
    Negative Logits
    ¿½
    -1.95
    °
    -1.85
    IJ
    -1.80
    för
    -1.75
    Ŀ
    -1.73
    ¦
    -1.71
    ı
    -1.64
    ¡
    -1.63
    ķ
    -1.62
    iors
    -1.58
    POSITIVE LOGITS
     love
    1.53
     planting
    1.44
     acids
    1.43
     Drain
    1.39
     plaster
    1.38
     plague
    1.36
     LSD
    1.31
     silk
    1.31
     psy
    1.31
     gelatin
    1.31
    Activations Density 1.815%

    No Known Activations