INDEX

Explanations

numerical quantities related to people and incidents

oai_token-act-pair · gpt-3.5-turbo

instances of the verb "to be" in various forms

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Relations

-0.76

 Â¯

-0.71

 Trey

-0.66

 Conclusion

-0.65

 srfAttach

-0.64

 Reply

-0.62

 Somew

-0.62

inventoryQuantity

-0.60

#$

-0.60

 Letter

-0.59

POSITIVE LOGITS

 affected

0.97

 crammed

0.94

 enrolled

0.94

 slaughtered

0.94

 killed

0.94

 surveyed

0.92

 mobilized

0.92

 registered

0.92

 impacted

0.90

 effected

0.88

Activations Density 0.113%

numerical quantities related to people and incidents

instances of the verb "to be" in various forms

No Comments

No Known Activations