INDEX

Explanations

phrases related to advising, providing information, or making suggestions

oai_token-act-pair · gpt-3.5-turbo

references to actions and capabilities

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 PsyNet

-0.66

worldly

-0.52

âģ

-0.52

awed

-0.51

Eh

-0.51

Um

-0.50

 Nope

-0.50

 Unloaded

-0.49

 Congratulations

-0.49

Well

-0.49

POSITIVE LOGITS

 also

2.04

 furthermore

1.77

also

1.66

 moreover

1.61

 additionally

1.55

 then

1.47

 ALSO

1.43

 likewise

1.33

 therefore

1.28

 Also

1.25

Activations Density 1.140%

phrases related to advising, providing information, or making suggestions

references to actions and capabilities

No Comments

No Known Activations