© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 4-RES-SM
    11. 10111
    Prev
    Next
    INDEX
    Explanations

    frequent usage of the word "widely."

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    78
    +0.11
    0.6%
    348
    +0.11
    0.6%
    72
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    429
    +0.11
    0.02
    65
    +0.11
    0.02
    72
    +0.11
    0.02
    Negative Logits
    Ļª
    -2.34
    ĺ
    -2.12
    ¿½
    -2.00
    ļ
    -1.94
    ľĵ
    -1.80
    į
    -1.77
    ĻĤ
    -1.77
    IJ
    -1.76
    ¦
    -1.74
    »¿
    -1.74
    POSITIVE LOGITS
     adopted
    2.30
     accepted
    2.25
     employed
    2.19
     available
    2.17
     deployed
    2.17
     used
    2.16
     regarded
    2.15
     studied
    2.15
     distributed
    2.10
     utilized
    2.07
    Activations Density 0.071%

    No Known Activations