© 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. 30043
    Prev
    Next
    INDEX
    Explanations

    instances of the word "lying" and its variations indicating deception or untruthfulness

    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
    46
    +0.14
    0.8%
    322
    +0.14
    0.8%
    255
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    46
    +0.14
    0.02
    322
    +0.14
    0.01
    352
    +0.11
    0.02
    Negative Logits
    wards
    -1.64
    BSD
    -1.46
    +)
    -1.38
    née
    -1.34
    rax
    -1.32
     hair
    -1.32
    +.
    -1.31
    agon
    -1.31
    aste
    -1.31
     gifts
    -1.30
    POSITIVE LOGITS
    uten
    2.10
    wers
    1.74
    hurst
    1.65
    utenant
    1.64
     dormant
    1.61
    offs
    1.54
    abouts
    1.53
    ering
    1.51
    led
    1.50
    eping
    1.49
    Activations Density 0.198%

    No Known Activations