INDEX

Explanations

references to the NSA (National Security Agency)

oai_token-act-pair · gpt-3.5-turbo

references to the NSA (National Security Agency)

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

sticks

-0.72

 paddle

-0.71

 clay

-0.71

Bott

-0.70

stick

-0.69

ustom

-0.66

Sty

-0.65

bb

-0.64

stone

-0.64

 Lanc

-0.62

POSITIVE LOGITS

NSA

3.66

NSA

3.25

 Snowden

1.99

 FISA

1.69

DEA

1.67

CIA

1.66

CHQ

1.53

TSA

1.52

FBI

1.48

CIA

1.44

Activations Density 0.016%

references to the NSA (National Security Agency)

references to the NSA (National Security Agency)

No Comments

No Known Activations