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

opian

-0.74

arts

-0.71

places

-0.67

©¶æ¥µ

-0.63

 press

-0.60

 Walton

-0.60

âĶģ

-0.59

ISH

-0.59

 Standards

-0.58

 Robertson

-0.58

POSITIVE LOGITS

hran

0.82

 risked

0.73

tc

0.70

arella

0.69

uci

0.69

berus

0.67

aret

0.66

é¾įå

0.65

emouth

0.65

arij

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.