INDEX

Explanations

phrases indicating sources or attributions

oai_token-act-pair · gpt-3.5-turbo

special characters and symbols

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 Franch

-0.93

 Nanto

-0.83

 snail

-0.81

 disse

-0.74

 Isle

-0.74

 Mayo

-0.71

 Lerner

-0.69

 stri

-0.68

 mainland

-0.68

 stamp

-0.67

POSITIVE LOGITS

Ŀ

1.80

¡

1.41

°

1.20

¦

1.19

©

1.15

ŀ

1.14

£

1.13

ª

1.12

ł

1.08

§

1.07

Activations Density 0.229%

phrases indicating sources or attributions

special characters and symbols

No Comments

No Known Activations