INDEX

Explanations

phrases related to systems or processes, especially those involving criticism or rigging

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 earnest

-0.76

azines

-0.71

tein

-0.70

 Emin

-0.66

 Dodgers

-0.64

joy

-0.63

 Aerial

-0.63

 Ranch

-0.62

terday

-0.62

 Padres

-0.61

POSITIVE LOGITS

atics

1.09

ically

0.94

wide

0.93

ologies

0.92

 overseen

0.85

 rigged

0.81

 governed

0.80

atically

0.78

opathy

0.78

ifice

0.77

Activations Density 0.090%

phrases related to systems or processes, especially those involving criticism or rigging

No Comments

No Known Activations