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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ible

-0.85

intent

-0.80

ainer

-0.78

uler

-0.75

itives

-0.72

enery

-0.69

meyer

-0.69

wered

-0.69

tumblr

-0.68

ief

-0.67

POSITIVE LOGITS

iT

0.72

 Gadget

0.69

theless

0.67

Ips

0.67

 Massacre

0.64

 Wave

0.64

 guiIcon

0.64

.>>

0.63

 Tsukuyomi

0.63

 Rooms

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.