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

    Neuronpedia

    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. 10999
    Prev
    Next
    INDEX
    Explanations

    abbreviations and acronyms related to organizations or classifications

    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
    369
    +0.13
    0.8%
    445
    +0.13
    0.8%
    213
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    410
    +0.13
    0.06
    445
    +0.13
    0.03
    369
    +0.13
    0.02
    Negative Logits
    abouts
    -1.74
     independence
    -1.54
    acia
    -1.45
     Expect
    -1.43
     competence
    -1.41
    hift
    -1.38
    nered
    -1.37
     ranging
    -1.37
     thou
    -1.36
     autonomy
    -1.30
    POSITIVE LOGITS
    ĨĴ
    2.27
    ±
    2.00
    arro
    1.97
    Ĵ
    1.95
    ŀ
    1.95
    «
    1.95
    Ń
    1.92
    etheless
    1.91
    ĻĤ
    1.89
    į
    1.85
    Activations Density 0.401%

    No Known Activations