INDEX

Explanations

proper nouns or names of individuals

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "had."

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

 Conclusion

-0.67

ellen

-0.65

Square

-0.64

 Mothers

-0.61

awa

-0.61

uating

-0.61

ickle

-0.61

coin

-0.60

DF

-0.60

FW

-0.59

POSITIVE LOGITS

 been

1.32

 begun

1.21

 gotten

1.20

 flown

1.17

 gone

1.12

 taken

1.10

 withdrawn

1.07

 undergone

1.07

 previously

1.06

 eaten

1.04

Activations Density 0.190%

proper nouns or names of individuals

occurrences of the word "had."

No Comments

No Known Activations