INDEX

Explanations

languages and language-related terms

oai_token-act-pair · gpt-3.5-turbo

references to languages and multilingual topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

kefeller

-0.78

apor

-0.76

romeda

-0.75

ramid

-0.75

arranted

-0.74

Reward

-0.73

apego

-0.73

 horm

-0.72

oppable

-0.71

rolet

-0.71

POSITIVE LOGITS

 languages

1.60

 language

1.53

language

1.49

 diction

1.49

 english

1.41

 Arabic

1.39

 English

1.37

english

1.34

English

1.34

 Hindi

1.33

Activations Density 0.665%

languages and language-related terms

references to languages and multilingual topics

No Comments

No Known Activations