INDEX

Explanations

phrases related to intelligence agencies and their conclusions on political matters

oai_token-act-pair · gpt-3.5-turbo

phrases related to assessments and conclusions made by intelligence officials regarding national security threats

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

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

MLA

-0.85

 Franchise

-0.82

 cunt

-0.80

 greedy

-0.80

 masc

-0.78

 labour

-0.78

 suff

-0.76

 Entreprene

-0.76

 Charity

-0.75

 labou

-0.74

POSITIVE LOGITS

 briefed

1.29

 counterterrorism

1.24

 intelligence

1.23

 Clapper

1.15

 Mattis

1.12

 declass

1.11

 Russia

1.10

 intercepted

1.10

 briefings

1.08

 intel

1.07

Activations Density 0.597%

phrases related to intelligence agencies and their conclusions on political matters

phrases related to assessments and conclusions made by intelligence officials regarding national security threats

No Comments

No Known Activations