INDEX

Explanations

phrases related to controversial topics or conflicts

oai_token-act-pair · gpt-3.5-turbo

mentions of potential or actual hotspots

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

 detach

-0.71

 Voices

-0.66

 hello

-0.65

 greeting

-0.64

 mourn

-0.63

 narrated

-0.62

 introduction

-0.60

Dri

-0.60

 Desmond

-0.60

 counseling

-0.60

POSITIVE LOGITS

pot

4.59

pots

3.59

Pot

2.45

pot

1.69

Pot

1.64

 pots

1.31

lot

1.19

hop

1.16

potion

1.09

stakes

1.09

Activations Density 0.011%

phrases related to controversial topics or conflicts

mentions of potential or actual hotspots

No Comments

No Known Activations