© 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. 28312
    Prev
    Next
    INDEX
    Explanations

    instances where clarification or elaboration is needed

    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
    356
    +0.13
    0.7%
    472
    +0.13
    0.7%
    258
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    472
    +0.13
    0.03
    148
    +0.13
    0.01
    327
    +0.11
    0.03
    Negative Logits
    ģ
    -1.84
    waukee
    -1.80
    eks
    -1.73
    ĨĴ
    -1.64
    ikes
    -1.62
    walks
    -1.61
    neh
    -1.61
    Ģ
    -1.61
    į
    -1.60
    nk
    -1.60
    POSITIVE LOGITS
     why
    2.69
     how
    2.21
     inconsistencies
    1.82
    why
    1.76
     discrepancies
    1.75
     gravity
    1.71
     discrepancy
    1.69
    how
    1.63
    mental
    1.60
     explanations
    1.58
    Activations Density 0.217%

    No Known Activations