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

    Neuronpedia

    Jacobian LensNEW
    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. 11081
    Prev
    Next
    INDEX
    Explanations

    various structural elements and attributes in HTML code

    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
    21
    +0.23
    1.4%
    78
    +0.23
    1.4%
    43
    +0.12
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    21
    +0.23
    0.04
    78
    +0.23
    0.01
    415
    +0.12
    0.04
    Negative Logits
    soever
    -1.65
    uality
    -1.58
     respectively
    -1.36
    GTH
    -1.34
     even
    -1.28
     described
    -1.26
     foreseeable
    -1.26
     included
    -1.25
    ayed
    -1.25
    aters
    -1.24
    POSITIVE LOGITS
    ³
    3.60
    ¶
    3.59
    §
    3.56
    į
    3.48
    «
    3.43
    ī
    3.41
    ²
    3.40
    ŀ
    3.40
    ĥ½
    3.39
    ©
    3.38
    Activations Density 0.161%

    No Known Activations