INDEX

Explanations

elements related to abstract concepts, including abbreviations and proper nouns

oai_token-act-pair · gpt-3.5-turbo

references to the abbreviation "AB" 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 @ 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

 Beir

-0.71

 explan

-0.70

 Maker

-0.65

 Disciple

-0.62

omore

-0.59

 Clause

-0.59

ancial

-0.59

Tid

-0.59

 Solitaire

-0.58

 virtue

-0.58

POSITIVE LOGITS

raham

1.52

stract

1.41

bey

1.24

ortion

1.18

dullah

1.17

oard

1.10

rams

1.07

bott

1.07

duct

1.02

bre

1.02

Activations Density 0.032%

elements related to abstract concepts, including abbreviations and proper nouns

references to the abbreviation "AB" in various contexts

No Comments

No Known Activations