INDEX

Explanations

large numbers

oai_token-act-pair · gpt-3.5-turbo

mentions of large quantities or numerical statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

eda

-0.76

anic

-0.76

 Dating

-0.72

GY

-0.72

bara

-0.71

 Counsel

-0.71

Boss

-0.70

ourage

-0.70

gradation

-0.70

 srfAttach

-0.68

POSITIVE LOGITS

 hectares

1.13

 tonnes

1.11

 dollars

1.11

 gallons

1.08

 acres

1.07

 euros

1.04

 inhabitants

1.03

 cubic

1.02

 Americans

1.02

 pounds

1.02

Activations Density 0.074%

large numbers

mentions of large quantities or numerical statistics

No Comments

No Known Activations