© 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. 12819
    Prev
    Next
    INDEX
    Explanations

    negative financial terms and significant figures related to financial obligations or budgets

    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
    17
    +0.14
    0.8%
    77
    +0.13
    0.7%
    197
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    77
    +0.14
    0.04
    197
    +0.13
    0.09
    391
    +0.13
    0.08
    Negative Logits
    ľĵ
    -2.19
    ĻĤ
    -2.06
    Į
    -1.88
    ulsive
    -1.72
    ĩ
    -1.64
    osure
    -1.63
    Ī
    -1.57
    numeric
    -1.57
    ¤
    -1.52
    unnable
    -1.49
    POSITIVE LOGITS
     County
    1.64
     cerevisiae
    1.57
     stake
    1.51
    consin
    1.51
     horiz
    1.44
     Hour
    1.40
     hereafter
    1.36
    sburg
    1.33
    apolis
    1.32
     Memorial
    1.32
    Activations Density 2.306%

    No Known Activations