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

orate

-0.93

Äĩ

-0.77

itability

-0.77

Operation

-0.71

 Normandy

-0.68

odan

-0.65

ttle

-0.65

 susceptibility

-0.64

rosis

-0.64

utation

-0.63

POSITIVE LOGITS

uel

0.70

nam

0.67

luaj

0.67

DSL

0.64

 contempl

0.64

 compr

0.64

taboola

0.64

mov

0.62

 entert

0.62

ãĤ¤ãĥĪ

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.