INDEX

Explanations

words related to emphasis or intensification

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing absolutes or strong modifiers

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

ategory

-0.72

senal

-0.72

DOM

-0.72

apons

-0.66

DragonMagazine

-0.65

toggle

-0.65

handle

-0.65

izons

-0.64

udi

-0.63

xit

-0.62

POSITIVE LOGITS

 uphill

0.89

 ironic

0.88

 unclear

0.84

 raining

0.81

 impossible

0.81

 evident

0.80

 obvious

0.75

 easy

0.74

ceivable

0.73

 conceivable

0.73

Activations Density 0.295%

words related to emphasis or intensification

phrases emphasizing absolutes or strong modifiers

No Comments

No Known Activations