INDEX

Explanations

phrases related to MI (likely referring to "Military Intelligence")

oai_token-act-pair · gpt-3.5-turbo

references to the UK intelligence agency MI6 and its operations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

scan

-0.87

itatively

-0.80

sw

-0.79

sworth

-0.77

flags

-0.75

immune

-0.71

ified

-0.70

markets

-0.70

marks

-0.70

ans

-0.70

POSITIVE LOGITS

xes

0.73

yre

0.72

enez

0.72

vine

0.72

olate

0.70

LEY

0.68

AMI

0.68

RE

0.68

lla

0.67

etta

0.67

Activations Density 0.162%

phrases related to MI (likely referring to "Military Intelligence")

references to the UK intelligence agency MI6 and its operations

No Comments

No Known Activations