© 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. 1-RES-SM
    11. 24161
    Prev
    Next
    INDEX
    Explanations

    structured data or code snippets

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    111
    +0.28
    1.8%
    271
    +0.26
    1.6%
    23
    +0.17
    1.1%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.28
    0.35
    271
    +0.26
    0.17
    156
    +0.17
    0.16
    Negative Logits
    hart
    -1.84
    eca
    -1.58
     dealership
    -1.51
    cias
    -1.50
    hardt
    -1.50
    izione
    -1.48
    asone
    -1.47
    ferent
    -1.46
    way
    -1.45
    cars
    -1.43
    POSITIVE LOGITS
     nothing
    1.74
     MERCHANTABILITY
    1.48
     everything
    1.45
    downarrow
    1.43
     God
    1.40
     eternity
    1.37
     deepest
    1.32
     him
    1.32
    being
    1.31
    cdots
    1.30
    Activations Density 0.477%

    No Known Activations