© 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. 1-RES-SM
    6. 27072
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "lies" and other forms related to falsehoods and deception

    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/1-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.1.hook_resid_post
    Hook Layer
    1
    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
    322
    +0.14
    0.8%
    476
    +0.13
    0.7%
    44
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    322
    +0.14
    0.02
    476
    +0.13
    0.03
    46
    +0.13
    0.02
    Negative Logits
    »¿
    -2.43
    Ŀ
    -2.39
    Ł
    -2.33
    ĺ
    -2.29
    Ķ
    -2.12
    Ħ
    -2.11
    Ģ
    -2.08
    Į
    -2.03
    ĭ
    -2.03
    Ĥ¬
    -2.00
    POSITIVE LOGITS
    uten
    1.84
    ental
    1.78
    ential
    1.59
    ational
    1.57
     dormant
    1.51
    inel
    1.44
    azole
    1.43
    linger
    1.43
    yond
    1.41
    zc
    1.40
    Activations Density 0.108%

    No Known Activations