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

ADVERTISEMENT

-0.66

ãģ®éŃĶ

-0.63

Downloadha

-0.63

==

-0.62

++)

-0.61

tes

-0.61

HAM

-0.61

teness

-0.60

ertodd

-0.60

Frames

-0.59

POSITIVE LOGITS

selves

0.75

inis

0.71

 Duty

0.69

 deductible

0.67

gow

0.67

zbollah

0.66

 disposition

0.64

ioned

0.64

 abund

0.62

riott

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.