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. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 7464
    Prev
    Next
    INDEX
    Explanations

    words related to political ideologies, particularly focusing on the term 'wing'

    oai_token-act-pair · gpt-3.5-turbo

    references to political ideologies, specifically left-wing themes

    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
     memor
    -0.67
     calendars
    -0.66
     Creat
    -0.63
     affirmative
    -0.62
     env
    -0.60
     recre
    -0.60
     depression
    -0.60
     rivalry
    -0.60
     mon
    -0.60
     Memor
    -0.59
    POSITIVE LOGITS
    wing
    4.67
    wings
    2.53
    Wing
    2.20
    leaning
    1.64
     wing
    1.59
     Wing
    1.49
    wed
    1.33
    ws
    1.24
    wered
    1.17
     wings
    1.14
    Activations Density 0.006%

    No Known Activations