INDEX

Explanations

mentions of time-related terms, specifically emphasizing the concept of 'early'

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "early" related to specific time periods

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

uyomi

-0.77

retched

-0.74

ortium

-0.73

leck

-0.69

erie

-0.68

ãĥ¯

-0.67

olulu

-0.67

shall

-0.66

ourge

-0.65

laws

-0.64

POSITIVE LOGITS

 early

0.89

 onset

0.78

 late

0.75

 mornings

0.74

 adop

0.73

 phases

0.73

 stages

0.73

 infancy

0.73

Early

0.72

 adolesc

0.71

Activations Density 0.024%

mentions of time-related terms, specifically emphasizing the concept of 'early'

occurrences of the word "early" related to specific time periods

No Comments

No Known Activations