INDEX

Explanations

mentions of legal and political events or actions

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

instances of the word "questions" and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

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

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_post

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

mobi

-0.70

ected

-0.69

rites

-0.68

 corrid

-0.66

Maps

-0.64

uj

-0.64

Chicken

-0.61

arbon

-0.60

olia

-0.59

 COUR

-0.58

POSITIVE LOGITS

naires

1.41

 posed

1.33

 answered

1.14

naire

1.13

 asked

1.11

 about

1.06

 pertaining

1.04

answered

1.03

iddles

0.98

 unanswered

0.96

Activations Density 0.096%

mentions of legal and political events or actions

instances of the word "questions" and related terms

No Comments

No Known Activations