INDEX

Explanations

instances where two entities are being compared or contrasted

oai_token-act-pair · gpt-3.5-turbo

instances where pairs of subjects or entities are mentioned together

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

Trend

-0.73

yip

-0.71

etter

-0.70

uel

-0.67

fecture

-0.65

assembly

-0.65

bor

-0.65

 Warden

-0.64

keleton

-0.63

inth

-0.63

POSITIVE LOGITS

edged

0.93

 equally

0.84

ocated

0.81

ocating

0.81

 respectively

0.80

 sides

0.77

 vying

0.75

 acknowledged

0.73

 denominations

0.72

ãĥīãĥ©ãĤ´ãĥ³

0.71

Activations Density 0.083%

instances where two entities are being compared or contrasted

instances where pairs of subjects or entities are mentioned together

No Comments

No Known Activations