INDEX

Explanations

terms or phrases in a context of dialogue

oai_token-act-pair · gpt-3.5-turbo

repeated special characters or symbols

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

 psychiat

-0.75

 destro

-0.72

 mosqu

-0.71

 Franch

-0.71

 Mayo

-0.70

mercial

-0.69

 representation

-0.66

 prevention

-0.65

 sequencing

-0.62

 fractures

-0.62

POSITIVE LOGITS

¹

1.26

ł

1.25

Ķ

1.24

¡

1.22

ª

1.21

Ŀ

1.18

ĳ

1.18

§

1.16

¦

1.15

į

1.14

Activations Density 0.151%

terms or phrases in a context of dialogue

repeated special characters or symbols

No Comments

No Known Activations