© 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. 31634
    Prev
    Next
    INDEX
    Explanations

    references to software versions and their features

    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
    98
    +0.15
    0.8%
    302
    +0.14
    0.8%
    136
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    457
    +0.15
    0.05
    362
    +0.14
    0.05
    75
    +0.12
    0.05
    Negative Logits
    jas
    -1.55
    oir
    -1.43
    zzi
    -1.40
     gust
    -1.40
    yrinth
    -1.37
     Applications
    -1.34
    oyer
    -1.34
     glimpse
    -1.33
     °
    -1.32
     degrees
    -1.31
    POSITIVE LOGITS
    ĨĴ
    2.08
    Īĺ
    2.02
    ĻĤ
    1.94
    ij
    1.83
    ľĵ
    1.77
    »¿
    1.76
    Ļª
    1.73
     usual
    1.72
    ¶
    1.61
     ours
    1.61
    Activations Density 0.408%

    No Known Activations