INDEX

Explanations

phrases indicating complete exclusion or avoidance

oai_token-act-pair · gpt-3.5-turbo

the word "altogether" 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 @ 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

 Patriarch

-0.71

 Cascade

-0.68

hao

-0.65

nan

-0.63

yers

-0.62

 Fulton

-0.62

yer

-0.62

 Apache

-0.61

Yar

-0.59

 Minute

-0.59

POSITIVE LOGITS

loo

0.79

 disarm

0.74

ADRA

0.68

olkien

0.65

ICA

0.65

gebra

0.65

EngineDebug

0.65

æ³

0.64

Iterator

0.64

Decre

0.64

Activations Density 0.010%

phrases indicating complete exclusion or avoidance

the word "altogether" in various contexts

No Comments

No Known Activations