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

ieties

-0.79

ITAL

-0.76

Ã£

-0.72

itals

-0.70

iens

-0.70

heses

-0.69

IRED

-0.69

¤

-0.68

itudes

-0.67

ikan

-0.66

POSITIVE LOGITS

 Ples

0.70

 compe

0.66

response

0.65

 Nicol

0.65

 response

0.64

Origin

0.64

 proced

0.64

 Orig

0.63

 Ping

0.61

 disg

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.