INDEX

Explanations

terms related to legislation, regulations, and policies

oai_token-act-pair · gpt-3.5-turbo

topics and keywords related to protections, education, labor, and scientific practices

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Parables

-0.71

é¾įå

-0.63

é¾į

-0.60

Klu

-0.59

£ı

-0.56

 Pixie

-0.55

 suffice

-0.55

 Guilty

-0.54

 Defin

-0.54

 Fortress

-0.53

POSITIVE LOGITS

0.74

etc

0.73

and

0.68

finder

0.67

etc

0.65

AND

0.62

rupted

0.61

or

0.59

itar

0.59

ADRA

0.59

Activations Density 0.390%

terms related to legislation, regulations, and policies

topics and keywords related to protections, education, labor, and scientific practices

No Comments

No Known Activations