INDEX

Explanations

numbers or percentages

oai_token-act-pair · gpt-3.5-turbo

numerical data, specifically percentages and statistics

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

zos

-0.74

SHIP

-0.73

{"

-0.72

abee

-0.71

>)

-0.70

Reading

-0.70

 Hawaiian

-0.70

favorite

-0.69

dirty

-0.68

Agent

-0.68

POSITIVE LOGITS

per

1.31

pc

1.23

 cent

1.09

 cents

1.08

pc

1.03

 manoeuv

1.03

 kilometres

1.02

 franc

1.01

 paed

0.99

 percentage

0.99

Activations Density 0.053%

numbers or percentages

numerical data, specifically percentages and statistics

No Comments

No Known Activations