INDEX

Explanations

adjectival phrases denoting a group of people

oai_token-act-pair · gpt-3.5-turbo

terms related to religious communities and gatherings

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

IFrame

Link

Negative Logits

Downloadha

-0.72

hran

-0.69

gery

-0.67

geries

-0.66

hya

-0.66

aundering

-0.66

PUT

-0.64

nsics

-0.63

ggle

-0.62

thin

-0.61

POSITIVE LOGITS

 congregation

1.08

 congreg

0.92

goers

0.88

 pastor

0.84

mates

0.83

 denominations

0.78

 members

0.77

rons

0.75

leader

0.75

arians

0.75

Activations Density 0.025%

adjectival phrases denoting a group of people

terms related to religious communities and gatherings

No Known Activations