© 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. 10-RES-JB
    11. 11100
    Prev
    Next
    INDEX
    Explanations

    words and phrases related to investigations, especially into political matters

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to investigations, particularly those involving political figures or incidents

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     partName
    -0.85
     ç¥ŀ
    -0.82
    tymology
    -0.76
    anguage
    -0.75
    ozyg
    -0.75
    ciation
    -0.74
    */(
    -0.74
     guiActiveUn
    -0.73
    retty
    -0.73
    terness
    -0.72
    POSITIVE LOGITS
     wrongdoing
    1.31
     misconduct
    1.25
     improper
    1.24
     corruption
    1.20
     collusion
    1.20
     abuses
    1.20
     mishand
    1.13
     malf
    1.12
     bribery
    1.11
     fraudulent
    1.10
    Activations Density 0.291%

    No Known Activations