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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

staking

-0.84

 encount

-0.83

 trave

-0.73

 nomine

-0.69

comed

-0.67

 misunder

-0.67

 unlaw

-0.66

 millenn

-0.66

hun

-0.64

 POLIT

-0.64

POSITIVE LOGITS

 stands

1.74

 boards

1.32

 runs

0.80

 nets

0.80

 sets

0.76

 prints

0.75

ockets

0.75

 strips

0.75

 ranks

0.74

 suites

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.