INDEX

Explanations

mentions of a specific person or character

oai_token-act-pair · gpt-3.5-turbo

non-English or special characters

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

letes

-0.71

cats

-0.67

 whales

-0.65

izard

-0.64

ukong

-0.63

Â¯

-0.63

 whale

-0.62

onian

-0.61

 Lizard

-0.60

bows

-0.59

POSITIVE LOGITS

 contention

1.11

leness

0.80

ãĥ¼ãĥĨ

0.79

 renegoti

0.75

illard

0.72

 sugg

0.70

FTA

0.69

ansion

0.65

 Shack

0.65

unin

0.65

Activations Density 0.000%

mentions of a specific person or character

non-English or special characters

No Comments

No Known Activations