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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

llo

-0.69

 grou

-0.65

ername

-0.63

 remark

-0.60

 comma

-0.60

 forbidden

-0.60

tem

-0.59

 prohibited

-0.58

 prank

-0.58

 spam

-0.57

POSITIVE LOGITS

redients

0.76

soDeliveryDate

0.74

acles

0.73

asta

0.67

ãĤ´ãĥ³

0.65

fixes

0.64

 ingred

0.63

abe

0.63

ecd

0.62

CD

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.