© 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. 30022
    Prev
    Next
    INDEX
    Explanations

    the term "little" 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
    338
    +0.14
    0.8%
    375
    +0.12
    0.6%
    198
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    372
    +0.14
    0.02
    338
    +0.12
    0.02
    375
    +0.11
    0.02
    Negative Logits
    Ļª
    -1.80
    ines
    -1.71
    roc
    -1.66
    ylase
    -1.64
    inated
    -1.61
    oder
    -1.58
    Ģ
    -1.57
    apore
    -1.56
    mlin
    -1.54
    IJ
    -1.54
    POSITIVE LOGITS
     bit
    1.71
     chance
    1.63
     resemblance
    1.62
     hearts
    1.53
    john
    1.52
    isy
    1.45
     eye
    1.43
     wonder
    1.43
     chunks
    1.40
     amount
    1.40
    Activations Density 0.100%

    No Known Activations