INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 thous

-0.68

 distances

-0.65

 rows

-0.65

Lin

-0.65

 curses

-0.63

 Downs

-0.63

 arenas

-0.62

Arcade

-0.61

 Pharma

-0.60

 resolutions

-0.60

POSITIVE LOGITS

haus

0.87

 somewhere

0.75

illac

0.75

leck

0.72

prising

0.71

sburgh

0.68

/.

0.67

 Prompt

0.64

zinski

0.64

 Courtesy

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.