INDEX

Explanations

instances where the word "fully" is used in the text

oai_token-act-pair · gpt-3.5-turbo

phrases indicating completeness or totality

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

 Guys

-0.77

 Disciple

-0.74

=-=-=-=-=-=-=-=-

-0.73

 Tycoon

-0.72

akov

-0.70

 Royale

-0.69

gee

-0.67

 foes

-0.66

 Spice

-0.66

 Springs

-0.66

POSITIVE LOGITS

 clothed

0.99

 qualified

0.91

 fled

0.88

qualified

0.86

 appreciated

0.85

 engaged

0.78

rehensive

0.76

fledged

0.76

charged

0.75

 utilized

0.75

Activations Density 0.018%

instances where the word "fully" is used in the text

phrases indicating completeness or totality

No Comments

No Known Activations