INDEX

Explanations

paragraphs in coding language

oai_token-act-pair · gpt-3.5-turbo

opening parentheses in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

EVs

-0.66

 Hemisphere

-0.65

ESV

-0.62

 fung

-0.62

{*

-0.60

 retali

-0.59

 nond

-0.58

TTL

-0.58

orate

-0.58

 Federation

-0.57

POSITIVE LOGITS

catentry

0.94

onduct

0.81

hello

0.77

aber

0.75

english

0.75

igration

0.75

antic

0.74

bour

0.73

outh

0.72

unicip

0.71

Activations Density 0.010%

paragraphs in coding language

opening parentheses in the text

No Comments

No Known Activations