© 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
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 2-RES-SM
    6. 12271
    Prev
    Next
    INDEX
    Explanations

    references to the quantity of items, especially in a context relating to music and technology

    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
    144
    +0.20
    1.1%
    166
    +0.13
    0.8%
    322
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    271
    +0.20
    -0.00
    144
    +0.13
    0.02
    66
    +0.13
    0.03
    Negative Logits
    verted
    -1.65
    vert
    -1.54
     planning
    -1.53
    elaide
    -1.51
     speaking
    -1.50
     manager
    -1.46
     stripping
    -1.39
    hered
    -1.38
     hiring
    -1.37
     favorable
    -1.37
    POSITIVE LOGITS
    ³
    2.53
    ¹
    2.32
    «
    2.27
    ¥
    2.26
    ª
    2.22
    °
    2.19
    ®
    2.16
    ·
    2.07
    Ī
    1.99
    Ļ
    1.96
    Activations Density 0.335%

    No Known Activations