INDEX

Explanations

news headlines or article titles ending with "Read more" links

oai_token-act-pair · gpt-3.5-turbo

high activation values indicating certain topics or themes related to news and current events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 mosqu

-0.71

 hooked

-0.69

 answ

-0.66

Ļ½

-0.66

 honored

-0.61

 homebrew

-0.60

XL

-0.59

mbuds

-0.59

 Compact

-0.59

scill

-0.58

POSITIVE LOGITS

than

0.87

Comments

0.77

perse

0.75

rug

0.70

dar

0.70

prev

0.69

fal

0.68

origin

0.67

notations

0.67

Fra

0.67

Activations Density 0.059%

news headlines or article titles ending with "Read more" links

high activation values indicating certain topics or themes related to news and current events

No Comments

No Known Activations