INDEX

Explanations

information related to intelligence agencies

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 bicycl

-0.85

galitarian

-0.80

Fest

-0.80

bernatorial

-0.79

 refunds

-0.77

Whe

-0.77

 ç¥ŀ

-0.76

 Blizz

-0.76

åħī

-0.76

 NASCAR

-0.75

POSITIVE LOGITS

 spies

1.83

 spying

1.83

spy

1.83

 espionage

1.82

 covert

1.66

KGB

1.66

 intelligence

1.64

CIA

1.59

CIA

1.58

ionage

1.52

Activations Density 0.832%

information related to intelligence agencies

No Comments

No Known Activations