INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

ctor

-0.73

 propagate

-0.73

atal

-0.68

acements

-0.66

 render

-0.65

ļé

-0.63

 caus

-0.63

hower

-0.62

 allocations

-0.61

 degrade

-0.61

POSITIVE LOGITS

flix

0.85

wcsstore

0.74

illin

0.71

Tang

0.68

oven

0.66

img

0.63

mode

0.63

Pil

0.63

wave

0.61

ctrl

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.