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

obin

-0.78

agle

-0.75

chev

-0.72

ram

-0.71

rists

-0.69

uph

-0.68

culus

-0.67

imm

-0.67

rontal

-0.66

ega

-0.65

POSITIVE LOGITS

eteria

0.77

Xie

0.73

Maker

0.65

 Aren

0.65

 Candle

0.61

 Qiao

0.61

aret

0.61

 Ying

0.61

dylib

0.61

Sut

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.