INDEX

Explanations

dates or references to specific time periods, especially the word "early."

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "early."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

def

-0.67

cro

-0.66

 itself

-0.65

 perm

-0.64

 bashing

-0.63

 rejo

-0.63

hog

-0.60

 suit

-0.60

 Belt

-0.59

 task

-0.59

POSITIVE LOGITS

early

4.56

Early

1.63

late

1.41

 Early

1.40

aturally

1.22

 early

1.18

umerous

1.08

Late

1.06

atively

1.02

EAR

1.00

Activations Density 0.024%

dates or references to specific time periods, especially the word "early."

occurrences of the word "early."

No Comments

No Known Activations