INDEX

Explanations

names of various individuals, potentially related to sports

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names of people and places

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 Gloria

-0.77

 Dorothy

-0.68

PBS

-0.66

 Wendy

-0.65

 Linda

-0.65

 Nobel

-0.65

 Watergate

-0.64

izoph

-0.63

 Karen

-0.63

RTX

-0.62

POSITIVE LOGITS

 struggled

1.14

 averaged

1.11

 underwent

1.08

 hasn

1.07

 possesses

1.06

 scored

1.05

 excel

1.03

 wears

1.03

 played

1.02

 suffers

1.00

Activations Density 0.203%

names of various individuals, potentially related to sports

proper nouns, particularly names of people and places

No Comments

No Known Activations