© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 26377
    Prev
    Next
    INDEX
    Explanations

    terms related to flexibility and physical exertion

    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
    +0.14
    0.8%
    394
    +0.14
    0.8%
    243
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    198
    +0.14
    0.02
    448
    +0.14
    0.01
    394
    +0.13
    0.03
    Negative Logits
     murdered
    -1.75
     chances
    -1.64
     dishon
    -1.63
     falsely
    -1.54
     killed
    -1.51
     deaths
    -1.50
     mention
    -1.50
     vain
    -1.50
    mur
    -1.49
    inois
    -1.47
    POSITIVE LOGITS
    ŀ
    3.35
    Ĵ
    3.13
    ĥ½
    3.07
    ķ
    3.07
    £
    3.03
    ĭ
    2.96
    Ŀ
    2.94
    ģ
    2.90
    Į
    2.86
    ľ
    2.83
    Activations Density 0.224%

    No Known Activations