INDEX

Explanations

instances of ongoing actions or processes

oai_token-act-pair · gpt-3.5-turbo

instances of the word "being" in various contexts

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

 endings

-0.84

 srfAttach

-0.73

 Elements

-0.67

 Awakens

-0.66

 ending

-0.65

 Newspaper

-0.64

 Closing

-0.62

 Beginning

-0.62

 Picks

-0.62

 conclusion

-0.62

POSITIVE LOGITS

 investigated

1.15

 eyed

1.13

 touted

1.12

 pursued

1.12

 chased

1.07

 hunted

1.05

 watched

1.02

 waged

1.02

amed

1.02

 phased

1.00

Activations Density 0.054%

instances of ongoing actions or processes

instances of the word "being" in various contexts

No Comments

No Known Activations