© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 2-RES-SM
    11. 10654
    Prev
    Next
    INDEX
    Explanations

    references to sample sizes in research contexts

    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/2-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.2.hook_resid_post
    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
    281
    +0.15
    0.8%
    100
    +0.14
    0.8%
    156
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    80
    +0.15
    0.02
    100
    +0.14
    0.02
    281
    +0.12
    0.01
    Negative Logits
    ŀ
    -3.07
    ĵ
    -2.94
    Ĵ
    -2.87
    ĨĴ
    -2.83
    §
    -2.77
    ĥ½
    -2.77
    Ł
    -2.69
    ĺ
    -2.66
    ĸ
    -2.61
    Ķ
    -2.60
    POSITIVE LOGITS
     sized
    1.90
     size
    1.81
     scra
    1.66
     grade
    1.61
     sizes
    1.52
     sample
    1.49
     drawn
    1.47
     samples
    1.43
     amounts
    1.42
     inexpensive
    1.42
    Activations Density 0.107%

    No Known Activations