© 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. 10107
    Prev
    Next
    INDEX
    Explanations

    references to the React JavaScript library and its components

    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
    156
    +0.30
    1.8%
    397
    +0.15
    0.9%
    64
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    38
    +0.30
    0.01
    369
    +0.15
    0.00
    321
    +0.11
    0.01
    Negative Logits
    ĥ½
    -2.04
    Ĭ
    -1.98
    į
    -1.86
    Ļª
    -1.86
    ¿½
    -1.80
    Ł
    -1.79
    Ĩ
    -1.73
    ¯
    -1.70
    thood
    -1.62
    ↵
    -1.62
    POSITIVE LOGITS
    olver
    1.92
    DOM
    1.73
    fiddle
    1.64
    ivable
    1.62
    namese
    1.54
    ERTY
    1.52
    master
    1.48
    esar
    1.45
    JS
    1.45
    uler
    1.42
    Activations Density 0.148%

    No Known Activations