© 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. 27905
    Prev
    Next
    INDEX
    Explanations

    formatted code and header references

    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
    478
    +0.13
    0.7%
    244
    +0.13
    0.7%
    410
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    487
    +0.13
    0.02
    315
    +0.13
    0.02
    192
    +0.13
    0.02
    Negative Logits
    manship
    -1.66
     masses
    -1.61
    pel
    -1.54
    elves
    -1.53
    wise
    -1.52
    plate
    -1.48
    protein
    -1.48
    like
    -1.48
    coins
    -1.46
    ably
    -1.40
    POSITIVE LOGITS
    »
    2.15
    ī
    1.97
    retval
    1.93
     mistaken
    1.92
    ¼
    1.86
    į
    1.84
     mistake
    1.64
    ·¸
    1.61
    µ
    1.61
    Į
    1.59
    Activations Density 0.141%

    No Known Activations