INDEX

Explanations

references to explosive devices, specifically grenades

oai_token-act-pair · gpt-3.5-turbo

references to grenades and grenade-related terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

tt

-0.81

itudes

-0.71

yll

-0.70

Madison

-0.68

Sax

-0.68

Penn

-0.68

Sea

-0.66

oney

-0.66

thereum

-0.65

ottage

-0.64

POSITIVE LOGITS

 grenades

1.23

 grenade

1.20

 launcher

1.07

adier

1.03

 Grenade

1.01

 launchers

0.98

 Launcher

0.98

 flares

0.95

 darts

0.80

 cans

0.79

Activations Density 0.012%

references to explosive devices, specifically grenades

references to grenades and grenade-related terminology

No Comments

No Known Activations