INDEX

Explanations

questions starting with "What do you think of" or "What do you think."

oai_token-act-pair · gpt-3.5-turbo

phrases that solicit opinions or feedback from the audience

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

atural

-0.71

ĸļ

-0.67

abad

-0.67

enges

-0.66

 Hels

-0.66

advertisement

-0.65

arget

-0.65

ylum

-0.64

Lago

-0.64

quad

-0.63

POSITIVE LOGITS

 guys

1.32

 think

1.11

 reckon

1.06

 prefer

1.05

 propose

0.97

 recommend

0.92

 want

0.90

 advise

0.88

 anticipate

0.88

 wanna

0.86

Activations Density 0.037%

questions starting with "What do you think of" or "What do you think."

phrases that solicit opinions or feedback from the audience

No Comments

No Known Activations