© 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. 4-RES-SM
    11. 11827
    Prev
    Next
    INDEX
    Explanations

    phrases that convey existence or state of being

    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/4-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.4.hook_resid_post
    Hook Layer
    4
    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
    466
    +0.21
    1.2%
    231
    +0.13
    0.7%
    454
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    23
    +0.21
    0.07
    144
    +0.13
    0.04
    402
    +0.12
    0.08
    Negative Logits
    ĻĤ
    -2.02
     whom
    -1.66
    [_
    -1.63
     and
    -1.63
    ]):
    -1.53
    ):
    -1.49
    esis
    -1.47
    ï¼ļ
    -1.43
    ():
    -1.40
    ?:
    -1.38
    POSITIVE LOGITS
     therefore
    2.37
     consequently
    2.29
     accordingly
    2.27
    therefore
    2.05
     thereafter
    1.93
     subsequently
    1.88
     likely
    1.80
     correspondingly
    1.77
     consequent
    1.75
     thus
    1.69
    Activations Density 0.732%

    No Known Activations