INDEX

Explanations

large numerical values likely referring to measurements

oai_token-act-pair · gpt-3.5-turbo

numerical values related to dimensions and financial figures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Parables

-0.69

 CASE

-0.62

 Extras

-0.61

 âĵĺ

-0.59

=(

-0.59

 Echoes

-0.58

PTS

-0.58

ogens

-0.58

 VIEW

-0.57

 Runes

-0.57

POSITIVE LOGITS

taboola

0.68

lling

0.67

 gallon

0.65

0.64

pole

0.63

shapeshifter

0.62

âĪĴ

0.57

pound

0.57

ounce

0.57

ailable

0.56

Activations Density 0.114%

large numerical values likely referring to measurements

numerical values related to dimensions and financial figures

No Comments

No Known Activations