INDEX

Explanations

words related to mocking or ridicule

oai_token-act-pair · gpt-3.5-turbo

instances of mockery or ridicule

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

enture

-0.83

 arrang

-0.83

 horizont

-0.78

ipeg

-0.75

inventoryQuantity

-0.73

 indic

-0.72

 cryst

-0.72

venture

-0.71

clerosis

-0.69

ccording

-0.67

POSITIVE LOGITS

ously

0.89

 mockery

0.86

 laughter

0.86

 ridicule

0.86

 mocking

0.84

 mocked

0.77

 scorn

0.74

fully

0.74

 ridic

0.71

 shouted

0.71

Activations Density 0.045%

words related to mocking or ridicule

instances of mockery or ridicule

No Comments

No Known Activations