© 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. 4-RES-SM
    11. 8569
    Prev
    Next
    INDEX
    Explanations

    possessive pronouns and references to personal ownership

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Neuron Alignment
    Index
    Value
    % of L₁
    21
    +0.18
    1.1%
    78
    +0.14
    0.8%
    509
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    3
    +0.18
    0.01
    338
    +0.14
    0.02
    10
    +0.12
    0.03
    Negative Logits
     fact
    -1.53
     certainty
    -1.53
     secrecy
    -1.49
    ishly
    -1.40
     bins
    -1.40
     templates
    -1.38
     void
    -1.37
     licenses
    -1.36
     detail
    -1.35
    fulness
    -1.35
    POSITIVE LOGITS
    Ń
    3.46
    ħ
    3.44
    IJ
    3.39
    ©
    3.28
    Ĩ
    3.20
    «
    3.02
    Ĺ
    2.96
    ij
    2.89
    »
    2.87
    Į
    2.85
    Activations Density 0.114%

    No Known Activations