© 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. 7-RES-JB
    11. 17111
    Prev
    Next
    INDEX
    Explanations

    information related to arrests and detainment

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to incarceration and legal proceedings

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 7-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    7
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Echoes
    -0.74
     nostalg
    -0.68
    ankind
    -0.66
    wolves
    -0.66
    phies
    -0.66
     VIDEOS
    -0.62
    DragonMagazine
    -0.62
    reddits
    -0.61
     Alps
    -0.60
    poll
    -0.60
    POSITIVE LOGITS
     parole
    1.25
     Detention
    1.17
     felony
    1.12
     detention
    1.11
     extradition
    1.11
     Correctional
    1.09
     jail
    1.07
     criminally
    1.07
     probation
    1.06
     arra
    1.06
    Activations Density 0.448%

    No Known Activations