INDEX

Explanations

time-related terms, particularly durations like 'weeks', 'months', and 'years'

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

references to specific time periods or durations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

oux

-0.75

rone

-0.75

onis

-0.70

alian

-0.70

rongh

-0.69

ropolitan

-0.69

inav

-0.68

ibrary

-0.67

tenance

-0.67

ynasty

-0.63

POSITIVE LOGITS

 preceding

1.22

 leading

1.19

 following

1.01

 ahead

1.00

 since

0.95

 immediately

0.94

 ensuing

0.92

 prior

0.91

 surrounding

0.88

 succeeding

0.84

Activations Density 0.098%

time-related terms, particularly durations like 'weeks', 'months', and 'years'

references to specific time periods or durations

No Comments

No Known Activations