© 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. 3092
    Prev
    Next
    INDEX
    Explanations

    references to individuals holding a significant position or title, possibly related to a specific institution or organization referred to as "CM."

    oai_token-act-pair · gpt-3.5-turbo

    mentions of the term "CM," which likely refers to a specific concept or entity in the context of the document

    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
    tein
    -0.82
    reys
    -0.81
    fell
    -0.76
    spell
    -0.74
    ggies
    -0.74
    lund
    -0.74
    rah
    -0.73
    gered
    -0.72
    ionage
    -0.69
    rets
    -0.69
    POSITIVE LOGITS
     Punk
    0.97
    ancel
    0.70
    oly
    0.68
    imes
    0.68
    orical
    0.66
    ory
    0.65
    Ģ
    0.65
    oglu
    0.63
    oral
    0.62
    oman
    0.62
    Activations Density 0.068%

    No Known Activations