© 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. 32303
    Prev
    Next
    INDEX
    Explanations

    high-level categorizations or labels

    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
    369
    +0.17
    1.0%
    87
    +0.14
    0.8%
    423
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    87
    +0.17
    0.03
    423
    +0.14
    0.04
    369
    +0.13
    0.02
    Negative Logits
    romycin
    -1.79
     gears
    -1.72
    ESULT
    -1.66
    ties
    -1.65
     intention
    -1.58
     breaks
    -1.48
    omitempty
    -1.44
    itary
    -1.41
     implementation
    -1.40
     outcome
    -1.35
    POSITIVE LOGITS
    ĸ
    2.49
    ķ
    2.42
    ircraft
    2.30
    Ŀ
    2.22
    thens
    2.22
    ureus
    2.22
    ĺ
    2.17
    «
    2.12
    »
    2.11
    ľ
    2.05
    Activations Density 0.104%

    No Known Activations