INDEX

Explanations

specific terms related to terrorist activities

oai_token-act-pair · gpt-3.5-turbo

mentions of hijackers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Rite

-0.80

âĨĳ

-0.73

 Angels

-0.72

remlin

-0.72

-0.70

 Apart

-0.67

ç¥ŀ

-0.66

mate

-0.63

 Blessed

-0.63

 Tiger

-0.63

POSITIVE LOGITS

hij

1.39

 hijacked

1.15

eering

1.03

intosh

0.83

inx

0.82

ackers

0.76

ibl

0.75

ulence

0.74

eer

0.73

querade

0.72

Activations Density 0.011%

specific terms related to terrorist activities

mentions of hijackers

No Comments

No Known Activations