INDEX

Explanations

proper nouns related to a specific character or place named Logan

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Logan."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ktop

-0.88

pse

-0.79

haps

-0.76

 dstg

-0.75

placed

-0.75

Prev

-0.72

ramid

-0.71

oulos

-0.70

tops

-0.69

dayName

-0.68

POSITIVE LOGITS

 Logan

1.31

 Lucky

0.84

 Airport

0.83

 airport

0.78

 Square

0.73

 Dull

0.72

McN

0.71

 Circle

0.71

istical

0.70

 Skywalker

0.70

Activations Density 0.007%

proper nouns related to a specific character or place named Logan

mentions of the name "Logan."

No Comments

No Known Activations