INDEX

Explanations

mentions of news publications and their coverage of various topics

oai_token-act-pair · gpt-3.5-turbo

references to prominent news sources

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

 subsistence

-0.70

weed

-0.69

unts

-0.67

ãħĭ

-0.64

animate

-0.63

olate

-0.63

 Takeru

-0.63

AFTA

-0.62

capacity

-0.61

 Viet

-0.61

POSITIVE LOGITS

 Editorial

0.91

 Investigative

0.90

 editorial

0.88

 quoting

0.84

 Transcript

0.84

 Newsp

0.83

Reader

0.82

antage

0.82

 headlined

0.78

 columnist

0.78

Activations Density 0.496%

mentions of news publications and their coverage of various topics

references to prominent news sources

No Comments

No Known Activations