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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

DIV

-0.80

NG

-0.75

oon

-0.71

Mo

-0.66

Via

-0.66

Magn

-0.65

inventoryQuantity

-0.62

POST

-0.61

Edited

-0.61

Loading

-0.61

POSITIVE LOGITS

abilia

0.72

 appellant

0.71

 behavi

0.71

unci

0.71

 demonstr

0.70

 exting

0.70

 Borough

0.70

etheless

0.68

ople

0.68

paio

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.