INDEX

Explanations

instances of the word "users" in text

oai_token-act-pair · gpt-3.5-turbo

references to users 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Hurricanes

-0.82

 Shore

-0.76

 Butcher

-0.74

western

-0.73

 Patriarch

-0.73

 Wolver

-0.72

 Baptist

-0.71

FLAG

-0.71

Rib

-0.70

 Scand

-0.69

POSITIVE LOGITS

pace

1.20

 users

1.19

 Users

0.99

paces

0.96

hip

0.93

wana

0.88

cript

0.88

 upgr

0.86

users

0.85

 passwords

0.82

Activations Density 0.020%

instances of the word "users" in text

references to users in various contexts

No Comments

No Known Activations