INDEX

Explanations

proper nouns related to individuals

oai_token-act-pair · gpt-3.5-turbo

subjects and pronouns in the context of opinions or actions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

population

-0.72

 Visitors

-0.69

assembly

-0.67

 Observatory

-0.63

 Equip

-0.62

 Ascension

-0.62

Statistics

-0.58

 Unicorn

-0.58

 mosqu

-0.58

 Canary

-0.58

POSITIVE LOGITS

 regretted

1.50

 reasoned

1.43

 disagreed

1.41

 didn

1.38

 knew

1.36

 objected

1.35

 felt

1.35

 argued

1.33

 wondered

1.31

 opted

1.29

Activations Density 0.542%

proper nouns related to individuals

subjects and pronouns in the context of opinions or actions

No Comments

No Known Activations