INDEX

Explanations

mentions of large numbers of people or users in different contexts

oai_token-act-pair · gpt-3.5-turbo

quantitative references to large groups of people or populations

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

odox

-0.87

imaru

-0.62

uncture

-0.61

 Compar

-0.61

SourceFile

-0.58

neau

-0.57

hetti

-0.57

 Dispatch

-0.57

 Whip

-0.57

 Rampage

-0.57

POSITIVE LOGITS

 worldwide

0.95

 annually

0.84

 nationwide

0.83

 acres

0.78

illion

0.77

 apiece

0.77

 crammed

0.76

pees

0.75

abytes

0.71

acht

0.70

Activations Density 0.259%

mentions of large numbers of people or users in different contexts

quantitative references to large groups of people or populations

No Comments

No Known Activations