INDEX

Explanations

proper nouns that include the letters "Sa" with high activations

oai_token-act-pair · gpt-3.5-turbo

mentions of individuals with the name "Saeed"

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

ounty

-0.76

emet

-0.75

 empath

-0.72

arks

-0.69

omen

-0.67

 tick

-0.66

rid

-0.65

yrinth

-0.65

ertodd

-0.64

odon

-0.63

POSITIVE LOGITS

Sa

3.82

Sa

3.06

sa

1.50

sa

1.37

Za

1.33

SA

1.26

Ma

1.18

Sah

1.18

 Sahara

1.17

Ta

1.17

Activations Density 0.014%

proper nouns that include the letters "Sa" with high activations

mentions of individuals with the name "Saeed"

No Comments

No Known Activations