INDEX

Explanations

proper nouns and specific terms related to political and governmental actions and entities

oai_token-act-pair · gpt-3.5-turbo

references to specific geographical locations or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

anwhile

-0.49

 southeastern

-0.45

 Supports

-0.45

taboola

-0.43

ashington

-0.43

arget

-0.42

formerly

-0.42

quartered

-0.41

heid

-0.40

 guiActiveUnfocused

-0.40

POSITIVE LOGITS

sag

0.53

 seiz

0.47

igans

0.47

 answ

0.45

 glim

0.44

 experien

0.41

 pige

0.41

girl

0.39

ãĤ§

0.39

papers

0.38

Activations Density 2.861%

proper nouns and specific terms related to political and governmental actions and entities

references to specific geographical locations or entities

No Comments

No Known Activations