INDEX

Explanations

phrases related to illegal activities, specifically bribery and extortion

oai_token-act-pair · gpt-3.5-turbo

terms related to bribery and corrupt practices

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

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

ãĥ´

-0.77

Cent

-0.76

ISS

-0.73

ember

-0.72

IST

-0.71

ART

-0.71

brew

-0.70

CONCLUS

-0.70

Adams

-0.69

noon

-0.69

POSITIVE LOGITS

 bribe

1.35

 bribes

1.26

 extortion

1.04

 bribery

0.93

 blackmail

0.91

 laundering

0.85

 disgu

0.85

 payments

0.84

 paid

0.83

 induce

0.82

Activations Density 0.018%

phrases related to illegal activities, specifically bribery and extortion

terms related to bribery and corrupt practices

No Comments

No Known Activations