INDEX

Explanations

mentions of surprising or shocking events or actions

oai_token-act-pair · gpt-3.5-turbo

words associated with surprise, shock, and unexpected outcomes

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

DragonMagazine

-0.64

rieve

-0.58

nings

-0.58

atana

-0.56

rosis

-0.56

rikes

-0.56

 regards

-0.56

uesday

-0.55

 defunct

-0.55

glas

-0.54

POSITIVE LOGITS

ively

0.97

 yourselves

0.87

 himself

0.82

ingly

0.81

 herself

0.80

 themselves

0.78

ilde

0.77

ibly

0.76

ously

0.76

fully

0.75

Activations Density 0.366%

mentions of surprising or shocking events or actions

words associated with surprise, shock, and unexpected outcomes

No Comments

No Known Activations