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

 waivers

-0.70

icons

-0.70

birds

-0.70

boats

-0.69

 wounding

-0.67

hedon

-0.66

chairs

-0.66

 prohibitions

-0.65

faced

-0.65

 dishon

-0.64

POSITIVE LOGITS

 neighbour

0.78

gd

0.70

 neigh

0.68

ppel

0.67

Gro

0.63

Ont

0.63

apolis

0.62

 Hazel

0.61

 Peel

0.61

gae

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.