© 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. 2-RES-SM
    11. 26505
    Prev
    Next
    INDEX
    Explanations

    technical jargon related to programming and data structures

    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
    419
    +0.35
    2.1%
    17
    +0.10
    0.6%
    263
    +0.09
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    156
    +0.35
    0.62
    111
    +0.10
    0.63
    71
    +0.09
    0.57
    Negative Logits
    IJ
    -2.81
    Ļª
    -2.74
    ¹
    -2.73
    £
    -2.72
    Ī
    -2.71
    Ģ
    -2.60
    ¼
    -2.58
    ĸ
    -2.55
    ļ
    -2.54
    ¸
    -2.51
    POSITIVE LOGITS
    occo
    1.57
    fficients
    1.47
    tone
    1.41
    rect
    1.40
    ature
    1.40
    hess
    1.39
     _________________________________
    1.38
    wagen
    1.38
    lace
    1.37
    erie
    1.36
    Activations Density 3.893%

    No Known Activations