INDEX

Explanations

mentions of news articles or press releases with the term "AP" (Associated Press)

oai_token-act-pair · gpt-3.5-turbo

mentions of the Associated Press (AP)

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

gone

-0.80

 capitalist

-0.68

abiding

-0.65

 ment

-0.65

stood

-0.64

 capitalists

-0.64

 occasion

-0.64

 gluten

-0.64

\(

-0.62

 Lords

-0.61

POSITIVE LOGITS

PLIED

1.10

osta

0.92

AP

0.91

PLIC

0.90

PAR

0.89

PLA

0.89

CBC

0.88

ENN

0.88

TN

0.86

PLE

0.83

Activations Density 0.011%

mentions of news articles or press releases with the term "AP" (Associated Press)

mentions of the Associated Press (AP)

No Comments

No Known Activations