INDEX

Explanations

numeric values mixed within text

oai_token-act-pair · gpt-3.5-turbo

elements related to technology and data representation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

aph

-0.73

 astroph

-0.73

 Marks

-0.71

 diagn

-0.69

 Flint

-0.68

Fay

-0.68

 Kens

-0.67

Ã©n

-0.67

 Beckham

-0.66

 Zheng

-0.66

POSITIVE LOGITS

1.30

1.14

ãĤ°

1.00

Â²

0.90

0.89

two

0.87

II

0.86

II

0.84

0.82

 Second

0.81

Activations Density 0.266%

numeric values mixed within text

elements related to technology and data representation

No Comments

No Known Activations