INDEX

Explanations

keywords related to surveys, news, opinions, or statistics

oai_token-act-pair · gpt-3.5-turbo

references to Americans and their perspectives or experiences

oai_token-act-pair · gpt-4o-mini Triggered 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

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

operation

-0.72

NB

-0.70

Initialized

-0.69

ival

-0.66

selection

-0.66

inventoryQuantity

-0.65

ÃŁ

-0.65

wrapper

-0.64

ijk

-0.63

fecture

-0.62

POSITIVE LOGITS

ICAN

0.82

 Idol

0.80

ornings

0.74

 Dollars

0.71

 citizens

0.70

who

0.70

 Airlines

0.69

 Skies

0.68

 born

0.68

 Citizens

0.68

Activations Density 0.032%

keywords related to surveys, news, opinions, or statistics

references to Americans and their perspectives or experiences

No Comments

No Known Activations