INDEX

Explanations

phrases related to medical care and ethics, particularly focusing on evidence-based practices

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

phrases that express skepticism or doubt regarding claims, qualifications, and the effectiveness or safety of products and policies

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

udeb

-0.80

 Worse

-0.73

ozo

-0.66

####

-0.66

ulla

-0.63

agh

-0.62

eting

-0.61

aceae

-0.60

Plot

-0.59

occupied

-0.59

POSITIVE LOGITS

 sincerity

1.73

 validity

1.48

worthiness

1.47

 reliability

1.43

 authenticity

1.41

 efficacy

1.38

 competence

1.38

 accuracy

1.37

 legitimacy

1.36

 adequ

1.36

Activations Density 0.317%

phrases related to medical care and ethics, particularly focusing on evidence-based practices

phrases that express skepticism or doubt regarding claims, qualifications, and the effectiveness or safety of products and policies

No Comments

No Known Activations