INDEX

Explanations

phrases involving contrasting or contradictory actions or situations

oai_token-act-pair · gpt-3.5-turbo

auxiliary verbs indicating states or conditions

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

acca

-0.88

kay

-0.80

eah

-0.75

resso

-0.74

oj

-0.70

minecraft

-0.69

rongh

-0.69

UD

-0.69

Mods

-0.68

onne

-0.67

POSITIVE LOGITS

 nonetheless

1.38

 nevertheless

1.22

etheless

0.87

 retains

0.83

 retaining

0.82

 retained

0.81

 acknow

0.80

 cautioned

0.78

 alas

0.76

 still

0.72

Activations Density 0.202%

phrases involving contrasting or contradictory actions or situations

auxiliary verbs indicating states or conditions

No Comments

No Known Activations