INDEX

Explanations

proper nouns related to a particular brand or company

oai_token-act-pair · gpt-3.5-turbo

references to specific names or titles in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

ernaut

-0.78

 tremend

-0.70

 bully

-0.65

 pets

-0.65

furt

-0.64

oway

-0.64

 reluct

-0.63

alogue

-0.60

facebook

-0.59

ascus

-0.59

POSITIVE LOGITS

ttes

0.81

Jub

0.71

 Hades

0.70

zza

0.69

este

0.67

 Thrust

0.67

 Cipher

0.66

 Throne

0.66

 Osiris

0.65

 Paradox

0.65

Activations Density 0.624%

proper nouns related to a particular brand or company

references to specific names or titles in the text

No Comments

No Known Activations