INDEX

Explanations

governmental and organizational terms related to designation and classification

oai_token-act-pair · gpt-3.5-turbo

terms related to political and societal classifications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 pregnancies

-0.69

scenes

-0.67

bats

-0.66

poons

-0.66

ãģĦ

-0.66

ousands

-0.65

abella

-0.63

ModLoader

-0.62

croft

-0.62

leys

-0.61

POSITIVE LOGITS

 worthy

1.08

 unto

1.05

 deserving

0.98

 akin

0.97

 capable

0.92

 requiring

0.89

 punishable

0.89

 devoid

0.86

 unlike

0.85

 suitable

0.83

Activations Density 0.278%

governmental and organizational terms related to designation and classification

terms related to political and societal classifications

No Comments

No Known Activations