INDEX

Explanations

phrases related to permission or authorization

oai_token-act-pair · gpt-3.5-turbo

phrases related to authorization and permission

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rir

-0.84

ï¸ı

-0.84

xual

-0.79

yip

-0.75

lycer

-0.72

edin

-0.71

 Emin

-0.71

 converge

-0.70

famous

-0.69

uably

-0.69

POSITIVE LOGITS

ISA

0.81

 authorizing

0.75

 decree

0.75

HCR

0.73

ADRA

0.72

 waiver

0.72

Applic

0.67

 behalf

0.67

 waivers

0.67

ĳå£«

0.66

Activations Density 0.318%

phrases related to permission or authorization

phrases related to authorization and permission

No Comments

No Known Activations