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

uana

-0.77

ilded

-0.77

iris

-0.73

spection

-0.72

regation

-0.71

viz

-0.66

eries

-0.66

abba

-0.66

glers

-0.65

liv

-0.64

POSITIVE LOGITS

 Introduced

0.72

Laun

0.64

 moratorium

0.64

 advis

0.63

cffffcc

0.63

 torped

0.62

 mant

0.62

umb

0.61

TION

0.60

 resil

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.