INDEX

Explanations

instances where there is a pause or interruption in speech or action

oai_token-act-pair · gpt-3.5-turbo

instances of pausing or interruptions in dialogue or action

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

CVE

-0.71

usted

-0.68

usky

-0.65

Assembly

-0.65

usting

-0.65

ipped

-0.64

allo

-0.63

çĭ

-0.62

gie

-0.61

ophile

-0.61

POSITIVE LOGITS

 pause

0.91

 pauses

0.82

gap

0.82

 periods

0.81

 paused

0.80

olicy

0.78

 momentarily

0.77

adian

0.76

 suspense

0.73

 indefinitely

0.71

Activations Density 0.079%

instances where there is a pause or interruption in speech or action

instances of pausing or interruptions in dialogue or action

No Comments

No Known Activations