INDEX

Explanations

words related to negative behavior or action

oai_token-act-pair · gpt-3.5-turbo

references to misdemeanors and their legal implications

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĥīãĥ©

-0.79

Pitt

-0.68

ulton

-0.67

ALT

-0.66

Crunch

-0.63

 subp

-0.62

HY

-0.62

éĹ

-0.62

benefit

-0.61

Downloadha

-0.61

POSITIVE LOGITS

ean

1.07

ours

0.99

ors

0.93

 Vaugh

0.92

els

0.82

omorph

0.81

eme

0.80

ements

0.76

acci

0.71

eman

0.71

Activations Density 0.007%

words related to negative behavior or action

references to misdemeanors and their legal implications

No Comments

No Known Activations