INDEX

Explanations

phrases related to taking action or making a statement

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "re."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

BuyableInstoreAndOnline

-0.76

 Pebble

-0.76

 lapt

-0.72

 Seah

-0.68

owl

-0.65

 Contra

-0.65

DERR

-0.61

Pom

-0.60

Koh

-0.60

 Marin

-0.59

POSITIVE LOGITS

aper

0.98

ngth

0.98

apers

0.95

versible

0.93

leases

0.89

becca

0.87

porter

0.83

claimer

0.82

ploy

0.81

ceive

0.81

Activations Density 0.040%

phrases related to taking action or making a statement

occurrences of the word "re."

No Comments

No Known Activations