INDEX

Explanations

expressions related to espionage or conspiracy

oai_token-act-pair · gpt-3.5-turbo

references to political and social manipulation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

"}],"

-0.83

uterte

-0.80

âĸº

-0.80

âĢ

-0.80

âĸł

-0.79

-0.78

âĻ¦

-0.78

Â»

-0.75

 Cosponsors

-0.74

 âĢ

-0.73

POSITIVE LOGITS

 basically

1.04

 crappy

0.98

 supposedly

0.96

 crap

0.96

 ridiculously

0.95

 magically

0.93

 downright

0.93

 hilar

0.91

 pretty

0.90

 clueless

0.89

Activations Density 1.071%

expressions related to espionage or conspiracy

references to political and social manipulation

No Comments

No Known Activations