© 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. 24631
    Prev
    Next
    INDEX
    Explanations

    expressions of humor and frustration

    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
    156
    +0.17
    0.9%
    27
    +0.13
    0.7%
    198
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    23
    +0.17
    0.16
    478
    +0.13
    -0.01
    366
    +0.13
    0.07
    Negative Logits
     differentially
    -1.65
     concurrently
    -1.49
     constit
    -1.38
     Constit
    -1.37
     addresses
    -1.34
    ï¬
    -1.33
    ubert
    -1.33
     advantage
    -1.32
     Oval
    -1.27
     impart
    -1.27
    POSITIVE LOGITS
    ¿½
    1.96
    letons
    1.90
     guy
    1.85
     doubt
    1.74
     didnt
    1.64
     dont
    1.56
     sorry
    1.51
    nt
    1.50
    ls
    1.49
    ½
    1.49
    Activations Density 4.534%

    No Known Activations