INDEX

Explanations

mentions of large quantities, specifically the word "tons"

oai_token-act-pair · gpt-3.5-turbo

references to large quantities or weights measured in tons

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

eve

-0.75

respect

-0.70

cher

-0.69

statement

-0.65

bank

-0.62

Ã§a

-0.62

Wilson

-0.62

 signalling

-0.62

iti

-0.62

Ã©e

-0.61

POSITIVE LOGITS

 tons

3.72

 tonnes

2.63

ton

1.89

 lots

1.59

 gallons

1.55

 loads

1.55

tons

1.49

 kilograms

1.45

 acres

1.38

 pounds

1.29

Activations Density 0.015%

mentions of large quantities, specifically the word "tons"

references to large quantities or weights measured in tons

No Comments

No Known Activations