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

 Seller

-0.71

 Calculator

-0.68

Haw

-0.67

 caller

-0.66

Owner

-0.65

Fee

-0.65

Dynamic

-0.63

rising

-0.62

IQ

-0.61

Max

-0.61

POSITIVE LOGITS

pmwiki

0.79

 Leban

0.78

 FranÃ§

0.73

Ire

0.72

0.71

thinkable

0.71

DEN

0.71

 foss

0.70

ospace

0.69

gans

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.