INDEX

Explanations

questions or prompts ending with a specific punctuation mark

oai_token-act-pair · gpt-3.5-turbo

parenthetical or list-like structures

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

closing parentheses in the text

oai_token-act-pair · gpt-4-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 guarding

-0.65

atron

-0.65

wagen

-0.63

 travers

-0.62

 continent

-0.62

 union

-0.62

 baskets

-0.61

 hust

-0.61

 truck

-0.61

 landslide

-0.60

POSITIVE LOGITS

 Continued

0.80

................................................................

0.78

Reviewer

0.78

 Hilbert

0.76

 CONTIN

0.75

HAM

0.74

 Enemies

0.73

 Conclusion

0.69

 Recommended

0.69

 References

0.68

Activations Density 0.336%

questions or prompts ending with a specific punctuation mark

parenthetical or list-like structures

closing parentheses in the text

No Comments

No Known Activations