INDEX

Explanations

titles or references related to a show or production

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals or groups, particularly in the context of closeness or relationships

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

encers

-0.83

 Investigators

-0.71

Reward

-0.64

 Wand

-0.62

XVI

-0.60

Staff

-0.60

 Forest

-0.60

Vit

-0.60

Mau

-0.59

rade

-0.58

POSITIVE LOGITS

oser

1.07

erker

0.87

 opio

0.79

 advoc

0.78

 horizont

0.78

uggest

0.77

ayer

0.74

amacare

0.73

cham

0.73

 entreprene

0.71

Activations Density 0.007%

titles or references related to a show or production

references to specific individuals or groups, particularly in the context of closeness or relationships

No Comments

No Known Activations