© 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. 4-RES-SM
    11. 9642
    Prev
    Next
    INDEX
    Explanations

    instances of self-referential phrases and the use of "they" related to opinions or observations about people

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    307
    +0.19
    1.1%
    274
    +0.18
    1.0%
    21
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    26
    +0.19
    0.12
    320
    +0.18
    0.11
    390
    +0.13
    0.10
    Negative Logits
     respectively
    -1.62
     ensued
    -1.62
     eds
    -1.59
     summary
    -1.56
    Discussion
    -1.52
     instructive
    -1.50
    umab
    -1.50
     farewell
    -1.49
     worth
    -1.47
     fruitful
    -1.46
    POSITIVE LOGITS
     already
    2.42
     only
    1.99
     never
    1.91
     lacks
    1.87
    already
    1.78
     hadn
    1.78
     didn
    1.70
     didnt
    1.69
     always
    1.69
     lacked
    1.65
    Activations Density 1.339%

    No Known Activations