INDEX

Explanations

instances of the word "former" followed by a related element, indicating a comparison or contrast between different entities

oai_token-act-pair · gpt-3.5-turbo

references to individuals or entities previously mentioned in a discourse

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

ammy

-0.83

otle

-0.81

atown

-0.79

antics

-0.78

andise

-0.75

hare

-0.75

ourced

-0.73

anders

-0.70

utterstock

-0.70

\\\\\\\\

-0.69

POSITIVE LOGITS

 Yugoslav

0.97

 Yugoslavia

0.89

est

0.83

 Soviet

0.74

 dictator

0.73

 smoker

0.71

 president

0.70

 communist

0.67

official

0.66

 mayor

0.65

Activations Density 0.032%

instances of the word "former" followed by a related element, indicating a comparison or contrast between different entities

references to individuals or entities previously mentioned in a discourse

No Comments

No Known Activations