© 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. 11-RES-JB
    11. 22560
    Prev
    Next
    INDEX
    Explanations

    proper nouns related to specific locations, authorities, and events in different countries

    oai_token-act-pair · gpt-3.5-turbo

    country demonyms and references to police forces or official entities

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Ó
    -0.85
    wn
    -0.77
    phe
    -0.76
    Õ
    -0.76
    same
    -0.74
    ||||
    -0.74
    isode
    -0.72
    reciation
    -0.68
    uggle
    -0.68
    vable
    -0.67
    POSITIVE LOGITS
     officials
    1.56
     lawmakers
    1.43
     authorities
    1.42
     prosecutors
    1.41
     investigators
    1.37
     commentators
    1.32
     politicians
    1.30
     regulators
    1.28
     observers
    1.27
     campaigners
    1.24
    Activations Density 0.183%

    No Known Activations