INDEX

Explanations

word parts related to time

oai_token-act-pair · gpt-3.5-turbo

references to time and existential reflections

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

tz

-0.64

 Previous

-0.61

xia

-0.60

historic

-0.59

ESE

-0.58

 Lens

-0.57

 BaseType

-0.57

 Tanzania

-0.56

Tir

-0.56

Orb

-0.55

POSITIVE LOGITS

 accordingly

1.18

 consequently

1.09

 alike

1.08

 thereafter

1.00

 thereof

0.88

 therein

0.86

 afterward

0.85

 afterwards

0.80

 therefore

0.77

 certainly

0.71

Activations Density 0.713%

word parts related to time

references to time and existential reflections

No Comments

No Known Activations