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

    Neuronpedia

    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. 11203
    Prev
    Next
    INDEX
    Explanations

    punctuation indicating significant pauses or transitions in thought

    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
    21
    +0.10
    0.6%
    223
    +0.10
    0.5%
    312
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    380
    +0.10
    0.00
    124
    +0.10
    0.00
    70
    +0.10
    0.00
    Negative Logits
     scene
    -1.67
    schemas
    -1.61
     reception
    -1.53
    åĬ¡
    -1.53
    orders
    -1.51
    hm
    -1.50
    ists
    -1.50
    ietz
    -1.50
     periphery
    -1.48
    omas
    -1.46
    POSITIVE LOGITS
    IJ
    2.08
    ĨĴ
    1.72
    ª
    1.54
    »¿
    1.48
     anonym
    1.44
    fitting
    1.42
    Ĵ
    1.36
    ŀ
    1.35
    lei
    1.35
    Ħ
    1.34
    Activations Density 0.000%

    No Known Activations