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

istrate

-0.71

Maps

-0.68

mus

-0.66

tested

-0.65

enne

-0.65

arine

-0.65

letters

-0.65

bug

-0.64

Natural

-0.63

Northern

-0.63

POSITIVE LOGITS

 rack

0.77

 reel

0.76

ãĤ¢ãĥ«

0.68

Alv

0.67

ãĥ«

0.66

 Kejriwal

0.65

 ç¥ŀ

0.64

 rented

0.64

0.63

"},"

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.