© 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. 13824
    Prev
    Next
    INDEX
    Explanations

    actions related to creativity and artistic expression

    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
    156
    +0.25
    1.4%
    186
    +0.16
    0.9%
    263
    +0.15
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    263
    +0.25
    0.12
    323
    +0.16
    0.06
    376
    +0.15
    0.02
    Negative Logits
    ®
    -4.43
    «
    -4.03
    Ĺ
    -3.99
    ↵
    -3.96
     ↵
    -3.96
    -3.96
           ↵
    -3.96
    ↵
    -3.96
    ↵
    -3.96
    ↵
    -3.96
    POSITIVE LOGITS
     yourself
    1.82
     freely
    1.78
     anymore
    1.78
     them
    1.67
     anyway
    1.55
     longer
    1.55
     ANY
    1.54
     sucks
    1.52
     either
    1.52
     whichever
    1.51
    Activations Density 1.281%

    No Known Activations