© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 30672
    Prev
    Next
    INDEX
    Explanations

    the word "As" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered 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
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    410
    +0.14
    0.8%
    424
    +0.13
    0.7%
    502
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    502
    +0.14
    0.06
    424
    +0.13
    0.04
    41
    +0.13
    0.05
    Negative Logits
    erent
    -1.64
    itud
    -1.57
    ered
    -1.53
    rained
    -1.47
    rance
    -1.47
     whom
    -1.43
     course
    -1.42
    orate
    -1.40
    UIT
    -1.38
    tees
    -1.38
    POSITIVE LOGITS
    ylum
    2.67
    ¶
    2.65
    ²
    2.59
    sembled
    2.56
    bestos
    2.39
    Ī
    2.37
    sembling
    2.34
    <|outofrange|>
    2.32
              ↵
    2.32
    2.32
    Activations Density 0.134%

    No Known Activations