© 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. 27014
    Prev
    Next
    INDEX
    Explanations

    scientific terms related to research methodologies and data analysis

    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
    21
    +0.17
    1.0%
    23
    +0.13
    0.7%
    450
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    21
    +0.17
    0.03
    236
    +0.13
    0.03
    251
    +0.10
    0.02
    Negative Logits
    ½
    -2.74
    ĥ½
    -2.72
    ĸ
    -2.70
    »¿
    -2.70
    Į
    -2.35
    Ī
    -2.34
    Ķ
    -2.26
    -2.25
    č↵
    -2.25
    -2.25
    POSITIVE LOGITS
    laid
    2.06
     manually
    2.05
     retrospectively
    1.97
     simultaneously
    1.87
     sequentially
    1.80
     separately
    1.77
     individually
    1.71
     surgically
    1.70
     successively
    1.67
     prospectively
    1.66
    Activations Density 0.139%

    No Known Activations