© 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. 6859
    Prev
    Next
    INDEX
    Explanations

    references to a specific news outlet, Politico

    oai_token-act-pair · gpt-3.5-turbo

    repeated references to the news outlet "Politico."

    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
    iour
    -0.76
     realised
    -0.74
     realise
    -0.70
     substitution
    -0.69
    master
    -0.68
     colours
    -0.67
     !!
    -0.66
     knees
    -0.65
     ts
    -0.64
    ievers
    -0.64
    POSITIVE LOGITS
     Politico
    3.14
     POLITICO
    2.98
     Newsweek
    2.17
     CNBC
    2.07
     BuzzFeed
    2.04
     Salon
    2.03
     Slate
    2.00
     Billboard
    1.96
     Gawker
    1.94
     HuffPost
    1.93
    Activations Density 0.059%

    No Known Activations