INDEX

Explanations

superlatives or phrases indicating high regard or excellence

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate rankings or superiority

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ption

-0.73

osponsors

-0.69

ples

-0.69

ittees

-0.67

ptoms

-0.65

mons

-0.64

parts

-0.61

rent

-0.58

lees

-0.58

onies

-0.57

POSITIVE LOGITS

 America

0.96

 Europe

0.95

 finest

0.86

 Britain

0.82

America

0.80

 humankind

0.80

 Asia

0.77

 mankind

0.77

 Scandinav

0.75

the

0.75

Activations Density 0.073%

superlatives or phrases indicating high regard or excellence

phrases that indicate rankings or superiority

No Comments

No Known Activations