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

 filler

-0.70

 extensions

-0.67

dam

-0.61

 compression

-0.60

 WRITE

-0.59

oline

-0.58

TPP

-0.58

corn

-0.57

semble

-0.57

Streamer

-0.57

POSITIVE LOGITS

lip

0.74

chio

0.68

Mos

0.65

 Hots

0.65

aciously

0.64

 desks

0.64

ewitness

0.64

atform

0.63

Cas

0.63

ï¸ı

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.