INDEX

Explanations

words related to communication or discussion

oai_token-act-pair · gpt-3.5-turbo

themes related to challenges and difficulties in communication and understanding

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

 Links

-0.59

 fateful

-0.58

interstitial

-0.56

 Chancellor

-0.53

 assorted

-0.53

mber

-0.51

 advisory

-0.51

 occasional

-0.51

 periodically

-0.51

MAY

-0.51

POSITIVE LOGITS

 without

1.65

without

1.52

 unless

1.47

 anymore

1.37

 WITHOUT

1.31

unless

1.19

 Without

1.13

 alone

1.00

 except

1.00

 because

0.99

Activations Density 0.608%

words related to communication or discussion

themes related to challenges and difficulties in communication and understanding

No Comments

No Known Activations