INDEX

Explanations

mentions of specific years and educational institutions

oai_token-act-pair · gpt-3.5-turbo

occurrences of specific symbols or characters in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ĵĺ

-0.82

omorphic

-0.77

conservancy

-0.73

ĸļ

-0.70

ĺħ

-0.69

 mathemat

-0.67

hon

-0.67

omnia

-0.66

ĻĤ

-0.64

wagen

-0.63

POSITIVE LOGITS

 Ibid

0.77

endif

0.61

 Sands

0.60

 Kislyak

0.59

quickShipAvailable

0.58

âĵĺ

0.58

NRS

0.58

Dialogue

0.57

 hers

0.57

addafi

0.57

Activations Density 0.602%

mentions of specific years and educational institutions

occurrences of specific symbols or characters in the text

No Comments

No Known Activations