© 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. 8634
    Prev
    Next
    INDEX
    Explanations

    first-person pronouns that indicate personal experiences or feelings

    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
    301
    +0.21
    1.2%
    498
    +0.14
    0.8%
    478
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    301
    +0.21
    0.20
    163
    +0.14
    0.15
    498
    +0.13
    0.15
    Negative Logits
    °
    -3.11
    Ħ
    -2.93
    º
    -2.81
    »¿
    -2.79
    ł
    -2.72
    ŀ
    -2.70
            ↵
    -2.70
    <|outofrange|>
    -2.70
     ↵
    -2.70
    <|outofrange|>
    -2.70
    POSITIVE LOGITS
    'm
    2.96
     am
    2.76
    've
    2.51
     believe
    2.50
     think
    2.48
     wish
    2.35
    brahim
    2.33
     hope
    2.30
     admire
    2.30
     confess
    2.26
    Activations Density 0.646%

    No Known Activations