INDEX

Explanations

phrases related to awards and titles, specifically "Player of the Year" recipients

oai_token-act-pair · gpt-3.5-turbo

mentions of awards, specifically "Player of the Year" titles across various sports

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

 streng

-0.74

sett

-0.73

pload

-0.71

 trave

-0.71

hett

-0.68

 ingred

-0.67

 arrang

-0.67

fram

-0.66

uala

-0.63

dor

-0.61

POSITIVE LOGITS

iverse

0.73

âĦ¢

0.66

 Awards

0.65

elight

0.63

 Mankind

0.62

Copyright

0.58

ago

0.55

cients

0.55

 ultr

0.54

worms

0.54

Activations Density 0.112%

phrases related to awards and titles, specifically "Player of the Year" recipients

mentions of awards, specifically "Player of the Year" titles across various sports

No Comments

No Known Activations