INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

iche

-0.70

verty

-0.65

 Vale

-0.63

 Battalion

-0.60

IV

-0.60

slave

-0.60

Lair

-0.59

Tro

-0.59

 Lair

-0.58

>,

-0.58

POSITIVE LOGITS

ALD

0.79

ãĤ¤ãĥĪ

0.69

amic

0.68

 hump

0.68

 Dems

0.67

 independents

0.65

imoto

0.65

izers

0.62

iting

0.62

rats

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.