© 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. 2-RES-SM
    11. 12778
    Prev
    Next
    INDEX
    Explanations

    symbols and formatting characters related to documents and coding

    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
    78
    +0.16
    0.9%
    122
    +0.13
    0.7%
    416
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    111
    +0.16
    0.26
    71
    +0.13
    0.27
    156
    +0.12
    0.27
    Negative Logits
     Wales
    -1.47
     mattered
    -1.46
     necessarily
    -1.45
     Welsh
    -1.42
    alent
    -1.41
     minded
    -1.38
    think
    -1.37
    wan
    -1.36
     Secondly
    -1.33
     cannot
    -1.29
    POSITIVE LOGITS
    ¦
    2.91
    Ŀ
    2.89
    ¼
    2.84
    ĭ
    2.80
    ¸
    2.74
    ¿½
    2.73
    İ
    2.72
    ĻĤ
    2.70
    Ļ
    2.68
    º
    2.61
    Activations Density 4.089%

    No Known Activations