INDEX

Explanations

phrases related to official statements or sources

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 charms

-0.72

gra

-0.67

;;;;;;;;;;;;

-0.66

Thumbnail

-0.66

 Cure

-0.65

 Rape

-0.64

 ç¥ŀ

-0.63

profits

-0.63

 Babe

-0.63

grass

-0.62

POSITIVE LOGITS

 unnamed

1.12

 source

1.01

Another

0.95

 sources

0.92

 unidentified

0.91

Sources

0.90

 insider

0.86

 Another

0.82

another

0.81

 personnel

0.80

Activations Density 0.262%

phrases related to official statements or sources

No Comments

No Known Activations