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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 1-RES-SM
    11. 28962
    Prev
    Next
    INDEX
    Explanations

    references to coffee and its consumption

    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
    +0.17
    1.0%
    173
    +0.11
    0.6%
    471
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    173
    +0.17
    0.02
    106
    +0.11
    0.02
    471
    +0.11
    0.02
    Negative Logits
    Ļª
    -2.25
    į
    -2.13
    ĺ
    -2.11
    ĻĤ
    -2.04
    Ŀ
    -1.94
    ĨĴ
    -1.87
    ĥ½
    -1.84
     transgender
    -1.83
    ij
    -1.77
    ľ
    -1.75
    POSITIVE LOGITS
    starter
    2.08
    shop
    1.93
    boat
    1.89
    houses
    1.81
    boats
    1.80
    holder
    1.74
    house
    1.74
     syrup
    1.70
    emaker
    1.70
    beans
    1.69
    Activations Density 0.122%

    No Known Activations