INDEX

Explanations

descriptions of attributes and capabilities of various products or objects

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing the qualities or characteristics of an object

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

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

inav

-0.83

ÄŁ

-0.76

ortmund

-0.76

uld

-0.74

osate

-0.73

ils

-0.72

cern

-0.72

ESE

-0.72

ends

-0.71

lag

-0.71

POSITIVE LOGITS

 capable

1.16

 versatile

1.13

 able

1.11

 prone

1.06

 extremely

1.03

 reminiscent

1.02

 flexible

1.01

 designed

1.01

 easy

1.00

 economical

0.98

Activations Density 0.505%

descriptions of attributes and capabilities of various products or objects

phrases emphasizing the qualities or characteristics of an object

No Comments

No Known Activations