INDEX

Explanations

Mentions of specific names, locations, and titles related to various individuals and places

oai_token-act-pair · gpt-3.5-turbo

geographic locations and names

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

OPLE

-0.62

staking

-0.58

 envy

-0.56

 puzz

-0.51

enegger

-0.45

WARE

-0.44

SHIP

-0.44

 neurot

-0.43

ACTED

-0.42

etheless

-0.41

POSITIVE LOGITS

oz

0.57

ani

0.53

uer

0.53

ali

0.53

ida

0.53

oka

0.53

am

0.52

alla

0.52

ona

0.52

aman

0.51

Activations Density 0.600%

Mentions of specific names, locations, and titles related to various individuals and places

geographic locations and names

No Comments

No Known Activations