INDEX

Explanations

specific keywords or phrases that are capitalized and often related to actions, events, or titles

oai_token-act-pair · gpt-3.5-turbo

prominent keywords or phrases that signify notable actions or subjects

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

 ..."

-0.61

.","

-0.54

\"

-0.53

''

-0.52

 ï¿½

-0.52

 thereto

-0.49

...

-0.49

.</

-0.48

``(

-0.47

 otherwise

-0.45

POSITIVE LOGITS

resa

0.90

odore

0.83

xiety

0.72

bidden

0.68

notations

0.68

tymology

0.68

jamin

0.67

ibliography

0.65

foundland

0.64

mosp

0.63

Activations Density 0.816%

specific keywords or phrases that are capitalized and often related to actions, events, or titles

prominent keywords or phrases that signify notable actions or subjects

No Comments

No Known Activations