INDEX

Explanations

mentions of organizations or proper nouns often followed by the letter 'G'

oai_token-act-pair · gpt-3.5-turbo

instances of the letter "G" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ĸļ

-0.86

 compr

-0.72

terday

-0.70

htt

-0.64

oÄŁ

-0.64

ĺħ

-0.63

 provision

-0.62

puter

-0.61

 retrospect

-0.59

 anonymously

-0.59

POSITIVE LOGITS

roups

1.24

iants

1.12

raphic

1.10

uild

1.10

ossip

1.04

reens

1.02

ambling

1.02

AMES

1.01

asp

0.99

rowth

0.98

Activations Density 0.033%

mentions of organizations or proper nouns often followed by the letter 'G'

instances of the letter "G" in various contexts

No Comments

No Known Activations