INDEX

Explanations

phrases related to locations, especially those including "LA" for Los Angeles

oai_token-act-pair · gpt-3.5-turbo

references to Los Angeles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 frost

-0.77

Mug

-0.70

Sug

-0.69

 deflation

-0.63

 wiser

-0.63

 bowl

-0.62

 crank

-0.61

 grades

-0.61

 coin

-0.61

 supervision

-0.60

POSITIVE LOGITS

LA

4.37

LA

1.77

LOS

1.52

CLA

1.51

Los

1.49

LM

1.43

LS

1.42

La

1.42

LI

1.41

la

1.38

Activations Density 0.011%

phrases related to locations, especially those including "LA" for Los Angeles

references to Los Angeles

No Comments

No Known Activations