© 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. 12878
    Prev
    Next
    INDEX
    Explanations

    URLs and web-related content with a focus on document structure and resource references

    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
    78
    +0.34
    2.5%
    397
    +0.18
    1.3%
    9
    +0.17
    1.3%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    9
    +0.34
    0.18
    78
    +0.18
    0.05
    397
    +0.17
    0.17
    Negative Logits
    than
    -1.52
    hs
    -1.27
    ino
    -1.24
    million
    -1.23
    hered
    -1.23
    inal
    -1.22
    erad
    -1.22
    uis
    -1.18
    expensive
    -1.16
    hering
    -1.15
    POSITIVE LOGITS
    ÑĨии
    1.67
    аÐ
    1.53
    оваÑĤÑĮ
    1.49
    ии
    1.48
    оÐ
    1.44
    ὶ
    1.41
     worden
    1.41
    া
    1.38
    ÑĨиÑı
    1.38
     werden
    1.35
    Activations Density 1.713%

    No Known Activations