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

 srfAttach

-0.71

orsche

-0.71

bright

-0.70

arching

-0.66

Ay

-0.65

MAT

-0.65

shell

-0.65

irlf

-0.64

ogly

-0.63

ricanes

-0.63

POSITIVE LOGITS

ulent

0.64

 futures

0.63

ãĥ¼ãĤ¯

0.62

ividual

0.61

ãĥł

0.61

ulence

0.60

emis

0.60

 Machina

0.59

reprene

0.59

 bargaining

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.