INDEX

Explanations

numeric time durations expressed in seconds

oai_token-act-pair · gpt-3.5-turbo

references to durations of time measured in seconds

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

hammer

-0.70

inav

-0.66

ju

-0.65

gat

-0.64

irie

-0.64

ãĥł

-0.64

iHUD

-0.63

annis

-0.63

gging

-0.63

 pilgr

-0.62

POSITIVE LOGITS

 elapsed

1.25

 intervals

0.90

ago

0.85

ngth

0.80

 cooldown

0.75

glass

0.75

 interval

0.75

 apiece

0.74

 increments

0.71

aneously

0.71

Activations Density 0.021%

numeric time durations expressed in seconds

references to durations of time measured in seconds

No Comments

No Known Activations