INDEX

Explanations

expressions of people's opinions or beliefs

oai_token-act-pair · gpt-3.5-turbo

phrases expressing conditionality or hypothetical scenarios involving the word "would."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

Kag

-0.67

 Tracker

-0.65

 Xuan

-0.64

mson

-0.59

 reminder

-0.56

 Compass

-0.55

{*

-0.55

 Brill

-0.55

rpm

-0.55

Ops

-0.55

POSITIVE LOGITS

 prefer

0.97

 gladly

0.93

rather

0.84

ĸļ

0.82

 otherwise

0.81

ivably

0.81

 ordinarily

0.79

 benefit

0.78

 willingly

0.77

 rather

0.75

Activations Density 0.240%

expressions of people's opinions or beliefs

phrases expressing conditionality or hypothetical scenarios involving the word "would."

No Comments

No Known Activations