INDEX

Explanations

personal pronouns and verbs indicating action or belief

oai_token-act-pair · gpt-3.5-turbo

phrases referencing individuals or groups identified by "who."

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

interstitial

-0.66

 srfAttach

-0.65

 Gothic

-0.64

 Apart

-0.63

 Delicious

-0.63

 Pets

-0.62

PokÃ©

-0.61

 Heights

-0.60

 Golem

-0.60

All

-0.59

POSITIVE LOGITS

 oppose

1.27

 criticize

1.12

 cherish

1.10

 specialize

1.09

 partake

1.09

 disagree

1.08

 believe

1.07

 participated

1.07

 wish

1.06

 dare

1.06

Activations Density 0.088%

personal pronouns and verbs indicating action or belief

phrases referencing individuals or groups identified by "who."

No Comments

No Known Activations