© 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. 30979
    Prev
    Next
    INDEX
    Explanations

    references to user account and identity management

    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
    450
    +0.16
    0.9%
    410
    +0.16
    0.9%
    17
    +0.15
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    17
    +0.16
    0.01
    450
    +0.16
    0.04
    379
    +0.15
    0.06
    Negative Logits
    ener
    -1.99
    iffer
    -1.76
    ening
    -1.69
    ienn
    -1.56
     thereof
    -1.55
    formed
    -1.47
    ashes
    -1.46
    unctional
    -1.45
    blockList
    -1.45
    urop
    -1.45
    POSITIVE LOGITS
    ¾
    3.35
    ¼
    2.95
    ½
    2.52
    Ģ
    2.46
    ¡
    2.40
    »
    2.36
    ¿
    2.34
    ģ
    2.34
    İ
    2.33
    Ľ
    2.33
    Activations Density 0.638%

    No Known Activations