© 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. 30075
    Prev
    Next
    INDEX
    Explanations

    parentheses and related punctuation in code or technical documentation

    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
    342
    +0.13
    0.7%
    98
    +0.11
    0.7%
    34
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    242
    +0.13
    0.02
    342
    +0.11
    0.02
    34
    +0.11
    0.02
    Negative Logits
    ®
    -1.71
    hydration
    -1.71
    Ĺ
    -1.49
    hoe
    -1.48
    assertEquals
    -1.46
    ovsk
    -1.46
    perm
    -1.44
     bladder
    -1.43
    »
    -1.42
    ľĵ
    -1.42
    POSITIVE LOGITS
    rant
    1.64
     Brussels
    1.54
     Gent
    1.51
     editors
    1.42
    oslov
    1.40
     certain
    1.39
     anonymous
    1.38
    geois
    1.36
     1981
    1.34
     1980
    1.32
    Activations Density 0.067%

    No Known Activations