© 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. 29861
    Prev
    Next
    INDEX
    Explanations

    references to remote access, control, or connection

    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
    156
    +0.24
    1.4%
    246
    +0.15
    0.9%
    438
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    246
    +0.24
    0.02
    304
    +0.15
    0.01
    30
    +0.11
    0.01
    Negative Logits
     radical
    -1.67
     bromide
    -1.62
    ophyll
    -1.56
    fully
    -1.51
     fibr
    -1.48
    ocrates
    -1.47
    askell
    -1.47
    oxygen
    -1.43
    ène
    -1.42
     rigorous
    -1.41
    POSITIVE LOGITS
     sensing
    1.82
    ÃŃo
    1.61
    pole
    1.58
    bie
    1.47
    FET
    1.46
    bat
    1.44
     (>
    1.41
    rowser
    1.40
    acks
    1.40
    chio
    1.40
    Activations Density 0.091%

    No Known Activations