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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

©¶æ

-0.73

zees

-0.73

sylv

-0.71

opsis

-0.69

 gorilla

-0.65

é¾įå

-0.62

Instance

-0.60

tle

-0.59

esan

-0.59

 Soviets

-0.58

POSITIVE LOGITS

externalActionCode

0.76

iage

0.72

anto

0.71

bands

0.66

ilot

0.66

 Hitch

0.66

ivo

0.66

urnal

0.64

acher

0.64

aning

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.