INDEX

Explanations

mentions of numerical or alphanumeric terms with negative values

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

negative sentiments or indications of decline

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

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

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_post

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

 survivors

-0.67

hon

-0.66

 internal

-0.63

 Nielsen

-0.61

 Ivanka

-0.61

 appeal

-0.59

 DeVos

-0.59

 Nigerian

-0.59

 provisional

-0.58

 conventional

-0.58

POSITIVE LOGITS

bats

1.67

bat

1.43

risk

1.34

large

1.33

times

1.26

home

1.22

capacity

1.10

sea

1.10

will

1.09

Large

1.07

Activations Density 0.014%

mentions of numerical or alphanumeric terms with negative values

negative sentiments or indications of decline

No Comments

No Known Activations