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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

shall

-0.83

thur

-0.72

fires

-0.68

itism

-0.66

bear

-0.65

ATURE

-0.61

 Athlet

-0.61

cot

-0.60

 Abbas

-0.60

 Equity

-0.58

POSITIVE LOGITS

 terms

1.51

way

0.85

chrom

0.74

retty

0.74

paralle

0.70

 Terms

0.67

©¶æ

0.65

tein

0.65

 ways

0.65

terms

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.