© 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. 31962
    Prev
    Next
    INDEX
    Explanations

    references to external sources or citations

    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
    369
    +0.17
    1.0%
    123
    +0.13
    0.7%
    444
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    460
    +0.17
    0.02
    324
    +0.13
    0.02
    314
    +0.13
    0.02
    Negative Logits
     startup
    -1.73
    "}](#
    -1.60
    doms
    -1.59
     shutdown
    -1.51
    ""
    -1.47
    ties
    -1.43
    upgrade
    -1.39
    acker
    -1.38
    ption
    -1.37
    autom
    -1.35
    POSITIVE LOGITS
    ĨĴ
    1.83
     Figures
    1.66
     Figure
    1.64
    ing
    1.57
     List
    1.56
    Ĵ
    1.52
    IJ
    1.52
    Ļª
    1.50
    boards
    1.50
    »¿
    1.50
    Activations Density 0.057%

    No Known Activations