INDEX

Explanations

phrases indicating multiple ways or reasons for something

oai_token-act-pair · gpt-3.5-turbo

phrases that suggest methods or reasons

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

livion

-0.82

 Tsukuyomi

-0.75

Â·Â·

-0.73

ĸļ

-0.70

ãģĻ

-0.64

Phones

-0.62

Mos

-0.60

UME

-0.60

onomic

-0.59

iture

-0.59

POSITIVE LOGITS

 pecul

0.75

 recurring

0.73

 inspir

0.72

 notable

0.71

 downside

0.68

 difference

0.66

 Redd

0.65

 highlights

0.64

 noteworthy

0.64

 redeem

0.64

Activations Density 0.148%

phrases indicating multiple ways or reasons for something

phrases that suggest methods or reasons

No Comments

No Known Activations