INDEX

Explanations

phrases indicating a problem or issue with something

oai_token-act-pair · gpt-3.5-turbo

phrases questioning the existence of problems or issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

mart

-0.88

swick

-0.76

eport

-0.67

lehem

-0.65

 Pesh

-0.64

pain

-0.63

onder

-0.63

grown

-0.62

izational

-0.62

 unden

-0.61

POSITIVE LOGITS

 regard

0.87

 regards

0.80

ĪĴ

0.78

 respect

0.71

 electrons

0.65

 Revenge

0.65

aroo

0.64

 dignity

0.62

otin

0.62

Ĥª

0.62

Activations Density 0.054%

phrases indicating a problem or issue with something

phrases questioning the existence of problems or issues

No Comments

No Known Activations