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

ificate

-0.70

 âĨ

-0.67

âĨ

-0.66

NPR

-0.65

 Atlas

-0.62

]]

-0.61

 anchor

-0.61

 transc

-0.60

âĹ¼

-0.60

pieces

-0.59

POSITIVE LOGITS

inki

0.75

unta

0.75

emi

0.68

ĪĴ

0.67

semble

0.67

asus

0.66

etting

0.66

 neighb

0.66

thritis

0.65

hell

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.