INDEX

Explanations

blades or blade-related words

oai_token-act-pair · gpt-3.5-turbo

references to blades, particularly in the context of tools or weapons

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

hett

-0.77

AMA

-0.73

LGBT

-0.72

 sitcom

-0.71

 Parks

-0.70

 Christians

-0.70

Privacy

-0.68

 LGBT

-0.67

sit

-0.67

aced

-0.66

POSITIVE LOGITS

 blade

3.42

 blades

3.24

blade

2.41

 Blades

2.20

Blade

2.19

 Blade

2.07

 sword

1.54

 rotor

1.51

 lightsaber

1.51

 knife

1.46

Activations Density 0.016%

blades or blade-related words

references to blades, particularly in the context of tools or weapons

No Comments

No Known Activations