INDEX

Explanations

references to additional information or related topics within a document

oai_token-act-pair · gpt-3.5-turbo

references or citations in a text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ufact

-0.76

idden

-0.68

iliate

-0.68

adoes

-0.68

adden

-0.63

athi

-0.63

 Cree

-0.62

ructose

-0.61

asc

-0.61

Delivery

-0.61

POSITIVE LOGITS

onyms

0.80

 Ùħ

0.79

...]

0.78

 âĿ

0.77

sein

0.77

Ù

0.73

ãģķ

0.71

 âĨ

0.70

nesses

0.69

 Advertisement

0.68

Activations Density 0.020%

references to additional information or related topics within a document

references or citations in a text

No Comments

No Known Activations