INDEX

Explanations

phrases indicating predictions or expectations

oai_token-act-pair · gpt-3.5-turbo

phrases indicating inevitability or certainty about future actions or outcomes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Enhancement

-0.67

cius

-0.65

lli

-0.59

Must

-0.59

 irresistible

-0.58

 largeDownload

-0.57

 methodological

-0.57

plates

-0.56

 motif

-0.56

 disclaimer

-0.56

POSITIVE LOGITS

 anymore

1.39

nor

1.00

 necessarily

1.00

 bother

0.99

 bothered

0.96

 anywhere

0.95

bud

0.92

any

0.88

swer

0.86

 EVER

0.86

Activations Density 0.231%

phrases indicating predictions or expectations

phrases indicating inevitability or certainty about future actions or outcomes

No Comments

No Known Activations