INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

TPPStreamerBot

-0.69

æĺ¯

-0.68

aley

-0.67

 lawy

-0.67

GV

-0.67

 streng

-0.66

ĸļ

-0.66

 tightening

-0.64

æĢ

-0.64

 å¤

-0.63

POSITIVE LOGITS

pha

0.71

cell

0.70

Mos

0.69

quel

0.68

iott

0.67

 EVENTS

0.63

 msec

0.62

Lua

0.61

DEN

0.61

strap

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.