© 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. 12691
    Prev
    Next
    INDEX
    Explanations

    elements related to design and aesthetics

    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
    189
    +0.17
    1.0%
    392
    +0.16
    0.9%
    37
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    189
    +0.17
    0.04
    392
    +0.16
    0.04
    360
    +0.11
    0.02
    Negative Logits
    §
    -1.96
    Ł
    -1.93
    ¯
    -1.91
    £
    -1.80
    Ľ
    -1.71
    Ĥ
    -1.70
    ĥ
    -1.69
    Ļª
    -1.66
    Ī
    -1.65
    chol
    -1.64
    POSITIVE LOGITS
    ists
    1.94
    ogue
    1.91
     design
    1.85
    ista
    1.76
     designers
    1.76
    emis
    1.75
    ement
    1.74
     rule
    1.69
    ators
    1.68
    igator
    1.67
    Activations Density 0.139%

    No Known Activations