© 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. 8-RES-JB
    11. 2688
    Prev
    Next
    INDEX
    Explanations

    words related to being aware or knowledgeable

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the suffix "iz" and similar forms that suggest the action or state of something

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    yip
    -0.72
     toget
    -0.61
    Unix
    -0.61
    jri
    -0.61
    IDS
    -0.59
    puter
    -0.59
    osate
    -0.59
    reddits
    -0.58
     Jinn
    -0.57
     grades
    -0.56
    POSITIVE LOGITS
    cess
    0.79
    ances
    0.74
    ional
    0.72
    ESS
    0.71
    å°Ĩ
    0.70
    itely
    0.70
    ance
    0.69
    æ³
    0.66
    amus
    0.65
    ripp
    0.64
    Activations Density 0.148%

    No Known Activations