INDEX

Explanations

phrases related to comparison, analysis, or evaluation

oai_token-act-pair · gpt-3.5-turbo

phrases indicating commonality or shared experiences among groups of people

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

SIGN

-0.69

hazard

-0.68

gression

-0.65

vertisement

-0.65

itud

-0.64

azard

-0.63

ernaut

-0.63

emption

-0.63

————————

-0.62

urden

-0.62

POSITIVE LOGITS

 chefs

1.04

 atheists

1.00

 astronomers

1.00

 rappers

0.97

 brewers

0.96

 feminists

0.96

 gamers

0.96

 philosophers

0.95

 actresses

0.93

 hikers

0.92

Activations Density 0.859%

phrases related to comparison, analysis, or evaluation

phrases indicating commonality or shared experiences among groups of people

No Comments

No Known Activations