INDEX

Explanations

phrases related to shocking or morally outrageous actions

oai_token-act-pair · gpt-3.5-turbo

expressions of shock or disbelief regarding human actions or decisions

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

-0.72

taboola

-0.69

Newsletter

-0.64

 respectively

-0.63

 Ghosts

-0.60

 Regist

-0.59

 Areas

-0.58

 Pigs

-0.58

 Trog

-0.57

 CrossRef

-0.57

POSITIVE LOGITS

 such

1.00

?!"

0.99

!?"

0.94

 seemingly

0.89

 blatantly

0.87

 mere

0.87

 suddenly

0.86

!?

0.81

?!

0.81

without

0.76

Activations Density 0.759%

phrases related to shocking or morally outrageous actions

expressions of shock or disbelief regarding human actions or decisions

No Comments

No Known Activations