INDEX

Explanations

products or entities that are being reviewed or evaluated

oai_token-act-pair · gpt-3.5-turbo

specific nouns or objects being referenced in the text

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

--------------------------------------------------------

-0.69

\<

-0.62

DERR

-0.62

itÃ©

-0.61

 Sections

-0.60

ween

-0.58

————————

-0.58

Unknown

-0.58

 foregoing

-0.57

CLR

-0.57

POSITIVE LOGITS

 belongs

0.81

 belonged

0.73

 represents

0.72

OULD

0.69

cture

0.67

 wont

0.66

 belong

0.65

eless

0.64

 lasted

0.63

 deserves

0.63

Activations Density 0.300%

products or entities that are being reviewed or evaluated

specific nouns or objects being referenced in the text

No Comments

No Known Activations