INDEX

Explanations

dates written as "une" followed by a number

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "une."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ories

-0.75

 teasp

-0.73

¶ħ

-0.73

EStreamFrame

-0.73

Ö¼

-0.72

loo

-0.71

İĭ

-0.70

lished

-0.70

BILITIES

-0.66

Â¢

-0.65

POSITIVE LOGITS

une

1.04

arthed

1.01

quist

0.69

 immunity

0.65

 insign

0.63

Mik

0.62

Yak

0.62

uce

0.60

CLASSIFIED

0.59

ices

0.59

Activations Density 0.008%

dates written as "une" followed by a number

occurrences of the word "une."

No Comments

No Known Activations