INDEX

Explanations

words indicating causation or consequence

oai_token-act-pair · gpt-3.5-turbo

the word "hence" and its variants, indicating causal relationships or explanations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

hea

-0.68

odcast

-0.62

 Featured

-0.62

 checklist

-0.58

attled

-0.55

Rev

-0.54

etsy

-0.54

 wrap

-0.54

 Wrap

-0.53

 dozen

-0.53

POSITIVE LOGITS

 hence

3.73

 thence

2.10

 consequently

2.06

 therefore

2.05

 Hence

2.02

 thus

1.85

 thereby

1.63

 furthermore

1.56

 whence

1.51

 moreover

1.47

Activations Density 0.011%

words indicating causation or consequence

the word "hence" and its variants, indicating causal relationships or explanations

No Comments

No Known Activations