INDEX

Explanations

phrases indicating speculation or assessment of likelihood

oai_token-act-pair · gpt-3.5-turbo

phrases that express uncertainty or subjective opinion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ilts

-0.79

zb

-0.67

onday

-0.66

estern

-0.62

ainers

-0.62

ffen

-0.61

ancock

-0.59

milo

-0.59

Peb

-0.58

ourse

-0.57

POSITIVE LOGITS

 like

1.00

 awfully

0.99

to

0.96

 innocuous

0.92

 tailor

0.87

 insur

0.87

 quaint

0.86

 destined

0.86

 unlikely

0.86

 harmless

0.83

Activations Density 0.066%

phrases indicating speculation or assessment of likelihood

phrases that express uncertainty or subjective opinion

No Comments

No Known Activations