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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

etheless

-0.75

cffffcc

-0.72

terday

-0.69

 downed

-0.69

 radar

-0.68

 Hurricanes

-0.64

 Predators

-0.64

romeda

-0.64

 acquaintance

-0.63

veyard

-0.62

POSITIVE LOGITS

lich

0.75

OH

0.72

oos

0.66

apter

0.65

heit

0.65

lance

0.65

buff

0.64

GD

0.64

 Amen

0.64

hyde

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.