INDEX

Explanations

phrases indicating a potential threat or danger

oai_token-act-pair · gpt-3.5-turbo

phrases indicating potential risks or threats

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

 Juice

-0.66

igger

-0.66

apolis

-0.63

 Weir

-0.62

cop

-0.60

 Fuller

-0.60

stream

-0.60

enum

-0.59

rew

-0.59

ln

-0.58

POSITIVE LOGITS

 pose

4.04

 poses

3.51

 posing

2.55

 posed

2.40

 Pose

1.86

 posture

1.41

pose

1.29

poses

1.19

 selfie

1.11

 gesture

1.02

Activations Density 0.024%

phrases indicating a potential threat or danger

phrases indicating potential risks or threats

No Comments

No Known Activations