INDEX

Explanations

news agency names

oai_token-act-pair · gpt-3.5-turbo

news agency names and reports

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

gregation

-0.73

Developer

-0.72

 Lovecraft

-0.72

artney

-0.67

 Trainer

-0.66

 Chevy

-0.66

etsy

-0.65

ophers

-0.65

 Purdue

-0.65

atown

-0.64

POSITIVE LOGITS

 newspaper

1.29

 tabloid

1.16

 newspapers

1.15

 newsp

1.11

aretz

1.10

ews

1.08

 news

1.04

 outlet

0.99

 daily

0.97

daily

0.97

Activations Density 0.313%

news agency names

news agency names and reports

No Comments

No Known Activations