INDEX

Explanations

keywords related to rankings or lists

oai_token-act-pair · gpt-3.5-turbo

ranks and positions in lists or classifications

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

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

uala

-0.75

igham

-0.74

ãĤ¶

-0.71

elcome

-0.69

issue

-0.69

pty

-0.69

ajor

-0.67

Ã§a

-0.67

adr

-0.66

Dialogue

-0.66

POSITIVE LOGITS

 earners

1.16

 contenders

1.05

 lists

1.01

 list

1.01

%:

0.99

 rankings

0.98

 spots

0.91

0.90

%,

0.89

 offenders

0.89

Activations Density 0.060%

keywords related to rankings or lists

ranks and positions in lists or classifications

No Comments

No Known Activations