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

iguous

-0.77

verning

-0.76

irection

-0.72

igible

-0.68

notations

-0.67

onga

-0.66

ACTED

-0.65

âķĲâķĲ

-0.65

orsi

-0.65

awaru

-0.65

POSITIVE LOGITS

stones

0.73

gren

0.71

elf

0.71

 Lonely

0.61

egg

0.60

archives

0.59

 anomalies

0.58

NEC

0.57

yea

0.57

plate

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.