© 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 EvalsExportsAPI 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. 10527
    Prev
    Next
    INDEX
    Explanations

    mentions of copyright

    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
    111
    +0.35
    2.1%
    156
    +0.33
    2.0%
    298
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    298
    +0.35
    0.02
    333
    +0.33
    0.01
    48
    +0.13
    0.01
    Negative Logits
    ¹
    -2.94
    ®
    -2.82
    Ł
    -2.79
    »
    -2.66
    ı
    -2.59
    Ĺ
    -2.52
    ↵  ³³³
    -2.49
    ↵
    -2.49
    -2.49
    -2.49
    POSITIVE LOGITS
     committees
    1.40
     behalf
    1.40
    icul
    1.38
     YE
    1.30
     Engineers
    1.27
    ÃŃc
    1.27
     Janeiro
    1.25
     Administration
    1.23
     AUTH
    1.23
     constitu
    1.21
    Activations Density 0.011%

    No Known Activations