INDEX

Explanations

text related to opting out or unsubscribing from something, specifically related to updates or classes

oai_token-act-pair · gpt-3.5-turbo

phrases related to subscription and opting out of communication

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

Pac

-0.92

pher

-0.78

omorphic

-0.78

matter

-0.76

ukong

-0.73

rats

-0.73

uably

-0.72

omon

-0.72

abal

-0.71

igans

-0.71

POSITIVE LOGITS

 email

0.85

 embed

0.79

 login

0.77

 subscribing

0.74

 emails

0.71

 rese

0.70

 notification

0.70

 Cancel

0.70

 Contact

0.70

 contact

0.69

Activations Density 0.042%

text related to opting out or unsubscribing from something, specifically related to updates or classes

phrases related to subscription and opting out of communication

No Comments

No Known Activations