INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

UTF

-0.83

feat

-0.75

âĹ¼

-0.75

utf

-0.73

ãĥ¼ãĥĨãĤ£

-0.73

udeb

-0.71

Um

-0.71

alg

-0.70

channelAvailability

-0.69

atro

-0.69

POSITIVE LOGITS

 offsets

0.69

 conferences

0.68

 submarines

0.68

isted

0.67

 defaults

0.67

 manuals

0.65

 luck

0.65

 talks

0.65

 ratios

0.65

 meetings

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.