INDEX

Explanations

No Explanations Found

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

)</

-0.74

worthiness

-0.70

Tube

-0.65

Avg

-0.61

UV

-0.61

edition

-0.61

hip

-0.61

ech

-0.59

eer

-0.58

Blu

-0.58

POSITIVE LOGITS

igham

0.79

 IMAGES

0.76

ipal

0.75

é¾įåĸļå£«

0.73

ertodd

0.73

neys

0.70

ãĤ¼ãĤ¦ãĤ¹

0.67

COL

0.66

eworthy

0.66

lords

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.