© 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. 28185
    Prev
    Next
    INDEX
    Explanations

    proper nouns, particularly names of people and locations

    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
    410
    +0.17
    1.0%
    466
    +0.16
    0.9%
    406
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    410
    +0.17
    0.04
    466
    +0.16
    0.03
    271
    +0.15
    -0.00
    Negative Logits
     otherwise
    -1.72
     worse
    -1.67
     late
    -1.60
     yet
    -1.55
    agine
    -1.55
    due
    -1.45
     acceptable
    -1.45
     larger
    -1.44
    ielle
    -1.44
    fitting
    -1.42
    POSITIVE LOGITS
    į
    2.15
    ĻĤ
    2.10
    º
    2.06
    ½
    2.03
    ETHOD
    1.97
    ´
    1.93
    ¼
    1.91
    ĭ
    1.91
    ¿
    1.89
    ľĵ
    1.88
    Activations Density 0.224%

    No Known Activations