© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    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. 9854
    Prev
    Next
    INDEX
    Explanations

    repeated mentions of the word "fact."

    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
    44
    +0.16
    0.9%
    221
    +0.12
    0.7%
    204
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    200
    +0.16
    0.03
    66
    +0.12
    0.03
    383
    +0.12
    0.03
    Negative Logits
    Īĺ
    -2.41
    Ħ
    -2.08
    »¿
    -1.98
    ĸ´
    -1.91
    ľ
    -1.89
    ¢
    -1.88
    ĭ
    -1.85
    į
    -1.84
    ĨĴ
    -1.82
    ¥
    -1.79
    POSITIVE LOGITS
    finding
    1.85
    itious
    1.82
    iko
    1.78
    oidal
    1.75
    uality
    1.74
    finder
    1.74
    awa
    1.74
    oid
    1.72
    sheet
    1.70
    IMO
    1.66
    Activations Density 0.267%

    No Known Activations