INDEX

Explanations

words related to activities involving mocking or making fun of something

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "bing" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 ORIG

-0.62

 Republic

-0.60

 honorable

-0.59

 Hemisphere

-0.58

 forwards

-0.57

 eras

-0.57

 approved

-0.57

 Miko

-0.56

 republic

-0.55

 accepted

-0.54

POSITIVE LOGITS

bing

4.42

bed

2.67

bers

2.13

ber

1.73

ilitating

1.61

bish

1.54

ming

1.51

ging

1.51

bling

1.49

bs

1.48

Activations Density 0.012%

words related to activities involving mocking or making fun of something

occurrences of the word "bing" in various contexts

No Comments

No Known Activations