© 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. 28210
    Prev
    Next
    INDEX
    Explanations

    words or terms related to "sand" 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
    156
    +0.18
    1.0%
    371
    +0.13
    0.7%
    18
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    371
    +0.18
    0.03
    18
    +0.13
    0.02
    304
    +0.11
    0.02
    Negative Logits
     direction
    -1.55
     neut
    -1.44
     pregnant
    -1.40
     ')
    -1.40
     subscrib
    -1.39
     sé
    -1.38
     breastfeeding
    -1.38
     consent
    -1.38
     particular
    -1.34
     seizure
    -1.32
    POSITIVE LOGITS
    bags
    2.18
    Ĵ
    2.11
    Ļª
    2.02
    ª
    2.01
    ±
    1.99
    ©
    1.99
    ĨĴ
    1.99
    ¯
    1.90
    ¢
    1.90
    bag
    1.89
    Activations Density 0.184%

    No Known Activations