INDEX

Explanations

words related to professional or technical terms

oai_token-act-pair · gpt-3.5-turbo

words that signify a pro or supportive stance towards various topics or issues

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

 Dwell

-0.73

 weave

-0.71

 gorge

-0.71

 Graveyard

-0.70

 EntityItem

-0.69

 ages

-0.69

 Slash

-0.68

 Halls

-0.67

 spears

-0.67

 Knife

-0.67

POSITIVE LOGITS

digy

1.13

secut

1.13

posal

1.11

gressive

1.08

secution

1.03

pro

1.01

blems

1.00

pos

0.98

bable

0.95

pose

0.94

Activations Density 0.009%

words related to professional or technical terms

words that signify a pro or supportive stance towards various topics or issues

No Comments

No Known Activations