INDEX

Explanations

phrases related to items or options that are diverse or varied

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "wide" 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 @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

aminer

-0.73

CB

-0.73

LV

-0.71

cair

-0.66

act

-0.65

ICAN

-0.65

 Maid

-0.65

ania

-0.64

nant

-0.64

ulia

-0.62

POSITIVE LOGITS

leaf

1.02

 wide

0.98

creen

0.94

 scope

0.91

 wider

0.89

 ranging

0.84

 widened

0.83

 breadth

0.83

 expans

0.81

 open

0.80

Activations Density 0.013%

phrases related to items or options that are diverse or varied

occurrences of the word "wide" in various contexts

No Comments

No Known Activations