INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

verse

-0.77

mes

-0.70

ippi

-0.69

gaard

-0.69

administ

-0.67

zhen

-0.66

nesses

-0.65

rome

-0.65

fold

-0.63

pdf

-0.63

POSITIVE LOGITS

obyl

0.78

 VIDEOS

0.74

 reper

0.67

licts

0.66

å£«

0.66

ĨĴ

0.65

 Liter

0.63

 rematch

0.62

levard

0.62

 erupted

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.