INDEX

Explanations

mentions of newspapers, specifically "The Wall Street Journal."

oai_token-act-pair · gpt-3.5-turbo

references to various news journals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ardless

-0.75

holders

-0.69

milo

-0.68

minus

-0.66

creen

-0.65

ailable

-0.64

 ours

-0.64

arding

-0.64

dh

-0.63

rav

-0.61

POSITIVE LOGITS

 Sentinel

1.04

 Journal

0.98

 Journals

0.86

ournals

0.86

Paper

0.84

 Editors

0.81

 Courier

0.81

 Papers

0.77

eers

0.77

Journal

0.76

Activations Density 0.008%

mentions of newspapers, specifically "The Wall Street Journal."

references to various news journals

No Comments

No Known Activations