INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 Beaver

-0.81

 Mons

-0.78

Oro

-0.77

 Mata

-0.73

 Noon

-0.72

 Defenders

-0.70

Ou

-0.67

 earthquakes

-0.67

 Yamato

-0.67

 observ

-0.65

POSITIVE LOGITS

ascar

0.70

paren

0.69

cca

0.69

GT

0.68

TC

0.68

alloc

0.67

yn

0.67

ecd

0.67

conservative

0.67

imb

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.