Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 2-RES-SM
    6. 28938
    Prev
    Next
    INDEX
    Explanations

    structured data representations and attributes associated with items like sizes and 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/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
    121
    +0.15
    0.9%
    143
    +0.12
    0.7%
    499
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    249
    +0.15
    0.06
    121
    +0.12
    0.08
    155
    +0.10
    0.15
    Negative Logits
     block
    -1.49
     etc
    -1.45
     ">
    -1.34
     processor
    -1.33
     storing
    -1.33
    sembles
    -1.27
    baum
    -1.27
    ew
    -1.26
     stored
    -1.25
     thumb
    -1.22
    POSITIVE LOGITS
    ¿½
    2.16
    ĥ½
    2.12
    ł
    1.91
    Į
    1.89
    Ī
    1.86
    ĩ
    1.85
    Ł
    1.83
    Ŀ
    1.75
    į
    1.70
    ī
    1.68
    Activations Density 5.611%

    No Known Activations