INDEX

Explanations

time-related terms with a focus on durations in months

oai_token-act-pair · gpt-3.5-turbo

mentions of time durations in months

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

anooga

-0.89

 Haram

-0.84

 hypocr

-0.76

otle

-0.74

oos

-0.74

Alias

-0.72

aque

-0.71

ocratic

-0.71

\\\\\\\\

-0.69

icket

-0.69

POSITIVE LOGITS

ago

0.98

pring

0.88

 Ukrain

0.87

 Ples

0.87

 pregnant

0.84

assetsadobe

0.75

Ago

0.73

nd

0.73

flower

0.72

ths

0.71

Activations Density 0.043%

time-related terms with a focus on durations in months

mentions of time durations in months

No Comments

No Known Activations