INDEX

Explanations

terms related to corruption

oai_token-act-pair · gpt-3.5-turbo

references to corruption and its related 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

nee

-0.83

imus

-0.80

leigh

-0.78

Â¯Â¯Â¯Â¯

-0.76

AY

-0.75

thur

-0.73

ocard

-0.71

cknowled

-0.71

liner

-0.71

ERSON

-0.70

POSITIVE LOGITS

 scandals

1.00

 corruption

0.92

 perpetrated

0.83

 corrupt

0.83

 schemes

0.82

 dealings

0.82

 disproportion

0.80

 endemic

0.80

 scandal

0.79

 rife

0.76

Activations Density 0.023%

terms related to corruption

references to corruption and its related contexts

No Comments

No Known Activations