INDEX

Explanations

contractions and words related to decision-making

oai_token-act-pair · gpt-3.5-turbo

positive expressions or affirmations about experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

But

-0.76

However

-0.73

But

-0.71

Ã±a

-0.70

but

-0.67

 However

-0.66

atl

-0.65

afa

-0.64

 WATCHED

-0.64

asp

-0.63

POSITIVE LOGITS

 nonetheless

1.80

 nevertheless

1.32

etheless

1.12

 darn

0.98

 awfully

0.93

 anyways

0.91

 anyway

0.87

 certainly

0.85

 damn

0.85

 gist

0.83

Activations Density 1.014%

contractions and words related to decision-making

positive expressions or affirmations about experiences

No Comments

No Known Activations