INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

oeuv

-0.63

 incrim

-0.61

pir

-0.61

TOR

-0.61

Sav

-0.61

ãĤª

-0.61

Player

-0.60

arming

-0.60

hep

-0.60

Iv

-0.58

POSITIVE LOGITS

picture

0.82

ilee

0.73

gee

0.73

Reviewed

0.70

re

0.69

kefeller

0.68

 Patch

0.67

 ribbon

0.67

taboola

0.66

 srfN

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.