© 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. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 0-RES-JB
    11. 2843
    Prev
    Next
    INDEX
    Explanations

    references to locations or groups that are predominantly of a certain type (e.g., ethnic, religious, gender)

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "predominantly" used in various contexts to describe demographics or distributions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    fax
    -0.72
    bugs
    -0.71
    opher
    -0.68
    bug
    -0.67
    Options
    -0.64
     decor
    -0.64
     Tours
    -0.64
    Master
    -0.62
     restores
    -0.61
    ebus
    -0.60
    POSITIVE LOGITS
     predominantly
    2.16
     disproportionately
    1.84
     predomin
    1.82
     disproportionate
    1.67
     minority
    1.66
    represented
    1.65
     minorities
    1.62
    majority
    1.55
     disproportion
    1.55
     Minority
    1.45
    Activations Density 0.027%

    No Known Activations