INDEX

Explanations

Roman numerals indicating sequential parts or sections within a larger document

oai_token-act-pair · gpt-3.5-turbo

references to parts or sequels of a series, as well as mentions of part-time work

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 pests

-0.68

ailability

-0.66

 receipt

-0.64

MSN

-0.61

Spider

-0.61

 scissors

-0.60

 pest

-0.59

Fal

-0.58

jan

-0.57

 instruments

-0.56

POSITIVE LOGITS

HDL

0.78

assium

0.73

ateral

0.72

otos

0.71

oyer

0.69

alion

0.68

lycer

0.67

iov

0.67

yss

0.67

ECA

0.66

Activations Density 0.125%

Roman numerals indicating sequential parts or sections within a larger document

references to parts or sequels of a series, as well as mentions of part-time work

No Comments

No Known Activations