INDEX

Explanations

phrases related to corporations or big entities

oai_token-act-pair · gpt-3.5-turbo

references to "big" in various contexts, often related to business and politics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 confir

-0.82

DragonMagazine

-0.77

yrinth

-0.74

 Cheong

-0.72

veyard

-0.72

 Machina

-0.71

theless

-0.70

cia

-0.70

qi

-0.69

inel

-0.68

POSITIVE LOGITS

oted

1.15

gest

1.03

wig

0.88

ger

0.86

ol

0.85

 chunk

0.85

GER

0.84

 bang

0.82

toe

0.80

mouth

0.78

Activations Density 0.040%

phrases related to corporations or big entities

references to "big" in various contexts, often related to business and politics

No Comments

No Known Activations