© 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. 11176
    Prev
    Next
    INDEX
    Explanations

    instances of conditional and purposeful language

    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
    507
    +0.14
    0.8%
    98
    +0.13
    0.7%
    478
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    407
    +0.14
    0.08
    282
    +0.13
    0.08
    238
    +0.13
    0.07
    Negative Logits
    hip
    -1.72
    hips
    -1.68
     dice
    -1.64
    ulence
    -1.62
     winds
    -1.54
    keit
    -1.50
    _[
    -1.46
    aux
    -1.44
    ?_
    -1.36
    ctomy
    -1.35
    POSITIVE LOGITS
              ↵
    3.56
    ↵↵
    3.56
    3.56
    č↵
    3.56
    ↵  âĢĥ
    3.56
    <|outofrange|>
    3.56
    ↵
    3.56
    ↵
    3.56
    ↵
    3.56
    3.56
    Activations Density 1.033%

    No Known Activations