INDEX

Explanations

mentions of legal charges or accusations against individuals

oai_token-act-pair · gpt-3.5-turbo

instances of the word "charged" in legal contexts

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

Wo

-0.78

birth

-0.73

COR

-0.70

 Birth

-0.68

ARCH

-0.66

 livest

-0.66

>]

-0.66

 Orth

-0.65

ophe

-0.64

angular

-0.63

POSITIVE LOGITS

heet

1.14

llah

0.93

 charges

0.91

eters

0.90

 criminally

0.80

 charging

0.78

 charged

0.77

hyde

0.76

 indict

0.76

Charg

0.73

Activations Density 0.028%

mentions of legal charges or accusations against individuals

instances of the word "charged" in legal contexts

No Comments

No Known Activations