© 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. 14158
    Prev
    Next
    INDEX
    Explanations

    verbs and nouns related to the impact or effect of actions on various entities

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to the effects and impacts of actions on health and society

    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
     commandments
    -0.76
     confess
    -0.70
     motto
    -0.66
     qualification
    -0.66
     Puzzles
    -0.65
     pursu
    -0.64
     desks
    -0.63
     mant
    -0.62
     begging
    -0.61
     Secrets
    -0.61
    POSITIVE LOGITS
     ripple
    1.20
     reverber
    1.13
    ffect
    1.09
    effects
    1.02
     negligible
    0.99
    Effects
    0.99
     impacts
    0.97
     effects
    0.96
     impact
    0.96
    impact
    0.95
    Activations Density 0.233%

    No Known Activations