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

eal

-0.84

olean

-0.76

ocative

-0.76

 Ibid

-0.74

usive

-0.72

earch

-0.69

amy

-0.67

emin

-0.67

evaluate

-0.66

atisf

-0.65

POSITIVE LOGITS

now

1.90

now

1.21

Now

1.12

NOW

1.11

Now

1.07

 still

0.73

 currently

0.72

NOW

0.72

 already

0.70

 presently

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.