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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 29968
    Prev
    Next
    INDEX
    Explanations

    expressions of surprise or excitement

    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
    369
    +0.16
    0.9%
    5
    +0.14
    0.8%
    144
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    13
    +0.16
    0.02
    410
    +0.14
    0.03
    420
    +0.13
    0.02
    Negative Logits
    CDATA
    -1.57
    encing
    -1.55
     backdrop
    -1.54
    emb
    -1.53
    rfloor
    -1.48
    ructure
    -1.48
    ency
    -1.46
    aterally
    -1.45
    iceps
    -1.43
    avorable
    -1.43
    POSITIVE LOGITS
    ammad
    1.78
     (@
    1.66
    lich
    1.60
    rame
    1.59
    ème
    1.53
     how
    1.47
    isms
    1.45
    len
    1.45
     speaking
    1.44
    ings
    1.42
    Activations Density 0.150%

    No Known Activations