INDEX

Explanations

specific references to medical or health-related terms

oai_token-act-pair · gpt-3.5-turbo

the definite article "the."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

rx

-0.61

wash

-0.61

ammed

-0.59

responsible

-0.55

olls

-0.55

 commits

-0.54

Fig

-0.54

berries

-0.53

wd

-0.53

oll

-0.53

POSITIVE LOGITS

 like

1.31

 occasional

1.27

 LIKE

1.05

 possibility

0.97

 ability

0.90

 inability

0.89

like

0.88

 consequ

0.87

 resultant

0.86

 afterlife

0.85

Activations Density 0.190%

specific references to medical or health-related terms

the definite article "the."

No Comments

No Known Activations