INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

pict

-0.74

Gum

-0.71

immune

-0.68

gypt

-0.66

Reviewer

-0.65

uitive

-0.62

-0.60

senal

-0.60

imov

-0.58

 Investigators

-0.58

POSITIVE LOGITS

ises

0.84

sung

0.73

ise

0.68

ize

0.65

ised

0.62

orest

0.61

orted

0.60

 sorts

0.59

 straw

0.59

heart

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.