INDEX

Explanations

phrases indicating exclusion or termination

oai_token-act-pair · gpt-3.5-turbo

the term "altogether" used in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

yers

-0.66

hma

-0.65

 Sims

-0.63

CCP

-0.62

nan

-0.62

agents

-0.61

 Agents

-0.61

DF

-0.61

UCS

-0.60

 Fulton

-0.59

POSITIVE LOGITS

 altogether

0.93

BuyableInstoreAndOnline

0.91

agher

0.83

 guiActiveUn

0.79

MpServer

0.78

ADRA

0.77

EngineDebug

0.74

 ç¥ŀ

0.73

olkien

0.72

DEV

0.72

Activations Density 0.004%

phrases indicating exclusion or termination

the term "altogether" used in various contexts

No Comments

No Known Activations