INDEX

Explanations

news agencies and related terms

oai_token-act-pair · gpt-3.5-turbo

mentions of news agencies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Founders

-0.71

 Courage

-0.68

 Generations

-0.63

atown

-0.63

 vanquished

-0.63

BIP

-0.62

 Solitaire

-0.62

competitive

-0.62

 Millennials

-0.60

ritch

-0.60

POSITIVE LOGITS

 agency

1.42

 outlet

1.36

wire

1.24

 portal

1.20

 agencies

1.12

 broadcaster

1.10

agency

1.09

 Agency

1.06

 outlets

1.06

 channel

1.04

Activations Density 0.046%

news agencies and related terms

mentions of news agencies

No Comments

No Known Activations