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

icken

-0.74

 mashed

-0.70

 MISS

-0.69

ALLY

-0.66

pause

-0.64

NetMessage

-0.64

taboola

-0.63

MID

-0.63

ymes

-0.63

ppo

-0.63

POSITIVE LOGITS

 degree

0.78

humane

0.71

eln

0.71

 elig

0.66

 Â¶

0.65

 inference

0.64

gnu

0.64

ktop

0.63

 Equality

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.