INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

letal

-0.81

Myth

-0.81

geist

-0.79

more

-0.77

thora

-0.75

mon

-0.72

olic

-0.71

æĪ¦

-0.71

ulty

-0.70

hot

-0.69

POSITIVE LOGITS

 capsule

0.72

 snapped

0.65

 cubes

0.64

 followed

0.63

 cube

0.63

 stripes

0.62

 follow

0.62

 photoc

0.62

CFR

0.61

 whisk

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.