INDEX

Explanations

dates or time references

oai_token-act-pair · gpt-3.5-turbo

instances of the word "dated" followed by numbers indicating specific dates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ffic

-0.82

utenberg

-0.79

ften

-0.78

yle

-0.77

Ross

-0.75

otos

-0.75

weeney

-0.73

akening

-0.73

IDES

-0.72

ËĪ

-0.72

POSITIVE LOGITS

 dated

1.13

dated

0.92

 Dating

0.84

 buggy

0.78

 Dates

0.75

 dates

0.72

cum

0.66

 dating

0.66

 date

0.66

ART

0.66

Activations Density 0.006%

dates or time references

instances of the word "dated" followed by numbers indicating specific dates

No Comments

No Known Activations