INDEX

Explanations

the word "Kam" with a numerical value following it

oai_token-act-pair · gpt-3.5-turbo

occurrences of the name "Kam" in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 rele

-0.68

 Phillies

-0.65

.ï¿½

-0.64

NRS

-0.64

 Scots

-0.64

ronic

-0.63

pull

-0.62

 predictive

-0.61

rec

-0.60

 corpus

-0.60

POSITIVE LOGITS

Kam

3.95

Hik

1.49

Kem

1.43

Kom

1.42

Kaw

1.41

Kum

1.41

Kap

1.41

Kas

1.41

Kal

1.34

Koz

1.34

Activations Density 0.027%

the word "Kam" with a numerical value following it

occurrences of the name "Kam" in various contexts

No Comments

No Known Activations