INDEX

Explanations

quantitative data such as estimated values and figures

oai_token-act-pair · gpt-3.5-turbo

numerical estimates related to measurements, values, or financial figures

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

mone

-0.88

ĪĴ

-0.84

warts

-0.71

 Flavoring

-0.70

redo

-0.64

Ħ¢

-0.61

 politic

-0.61

azard

-0.59

 distraction

-0.59

oran

-0.59

POSITIVE LOGITS

 ranges

1.05

 ranging

1.05

 averaging

1.04

 averaged

1.00

 range

1.00

 costing

0.97

 averages

0.96

 totaled

0.96

 estimated

0.94

 approx

0.93

Activations Density 0.447%

quantitative data such as estimated values and figures

numerical estimates related to measurements, values, or financial figures

No Comments

No Known Activations