© 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. 14155
    Prev
    Next
    INDEX
    Explanations

    a variety of key-value pairs and string fields in structured data or code snippets

    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
    478
    +0.17
    1.0%
    369
    +0.15
    0.9%
    261
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    261
    +0.17
    0.04
    388
    +0.15
    0.05
    466
    +0.14
    0.04
    Negative Logits
     existed
    -1.75
     emerged
    -1.60
     happened
    -1.60
     measures
    -1.52
     exists
    -1.51
    rays
    -1.48
    irds
    -1.46
     holds
    -1.44
     cannot
    -1.43
     survived
    -1.42
    POSITIVE LOGITS
    Ĵ
    3.91
    ĭ
    3.76
    Ĥ¬
    3.70
    ¦
    3.67
    ĥ½
    3.62
    Ĥ
    3.61
    <|outofrange|>
    3.57
    ↵        ↵
    3.57
    ↵↵
    3.57
    ↵
    3.57
    Activations Density 0.204%

    No Known Activations