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

ç¥ŀ

-0.71

 Learns

-0.69

é¾įåĸļå£«

-0.64

 Telescope

-0.61

 Racial

-0.61

 Learning

-0.61

ĪĴ

-0.58

 Lawyers

-0.58

serving

-0.58

ciating

-0.57

POSITIVE LOGITS

 Sabha

0.84

archment

0.84

bsite

0.73

raq

0.73

isks

0.67

ym

0.67

aus

0.66

amph

0.66

onom

0.66

ise

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.