© 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. 14268
    Prev
    Next
    INDEX
    Explanations

    terms related to participation or involvement in activities

    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/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
    Neuron Alignment
    Index
    Value
    % of L₁
    343
    +0.16
    0.9%
    165
    +0.12
    0.7%
    338
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    343
    +0.16
    0.04
    372
    +0.12
    0.02
    463
    +0.11
    0.02
    Negative Logits
    oths
    -1.61
    blogger
    -1.50
    ucker
    -1.45
    caster
    -1.43
     Minutes
    -1.42
    herty
    -1.42
    mouth
    -1.40
    worm
    -1.34
    rends
    -1.34
     tune
    -1.32
    POSITIVE LOGITS
    ł
    1.91
    ·
    1.87
    Ľ
    1.74
    ¹
    1.70
    asis
    1.69
    Ŀ
    1.66
    ľĵ
    1.66
    ment
    1.62
    ķ
    1.60
    ¬
    1.60
    Activations Density 0.168%

    No Known Activations