INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

iru

-0.87

atron

-0.82

bage

-0.81

ugal

-0.80

alo

-0.80

rio

-0.75

rils

-0.75

udicrous

-0.72

atan

-0.71

bol

-0.69

POSITIVE LOGITS

Thirty

0.63

 antagonists

0.63

displayText

0.61

 consent

0.60

 occup

0.60

QUEST

0.60

 chees

0.59

 nons

0.58

References

0.58

History

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.